Competitive Electricity Markets: Design, Implementation, Performance (Elsevier Global Energy Policy and Economics Series)

EBL Competitive Electricity Markets: Design, Implementation, Performance EBL EBL This page intentionally left blan...

Author: Fereidoon Perry Sioshansi

180 downloads 1523 Views 5MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD

EBL

Competitive Electricity Markets: Design, Implementation, Performance

EBL

EBL

This page intentionally left blank

Competitive Electricity Markets: Design, Implementation, Performance Edited by Fereidoon P. Sioshansi Menlo Energy Economics 1925 Nero CT Walnut Creek, CA

EBL

AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD PARIS • SAN DIEGO • SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO

Elsevier The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands First edition 2008 Copyright © 2008 Elsevier Ltd. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress

EBL

ISBN: 978-0-08-047172-3

For information on all Elsevier publications visit our website at books.elsevier.com Printed and bound in Great Britain 08

09

10

11

12

10

9

8

7

6

5

4

3 2

1

Working together to grow libraries in developing countries www.elsevier.com | www.bookaid.org | www.sabre.org

Contents Contributors

vii

Foreword: Liberalization and Regulation in Electricity Systems – How can We get the Balance Right? Michael Pollitt Preface: Competition and Long-Term Dimensions of Electricity Supply Wolfgang Pfaffenberger Introduction: Electricity Market Reform – Progress and Remaining Challenges Fereidoon P. Sioshansi

2.

3.

xxxv

1

25

PART I: Market Reform Evolution 1.

xvii

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets Hung-Po Chao, Shmuel Oren, and Robert Wilson

EBL

Hybrid Electricity Markets: The Problem of Explaining Different Patterns of Restructuring A.F. Correljé and L.J. De Vries Achieving Electricity Market Integration in Europe Nigel Cornwall

PART II: Market Performance, Monitoring and Demand Participation 4.

Transmission Markets, Congestion Management, and Investment Harry Singh

5.

The Design of US Wholesale Energy and Ancillary Service Auction Markets: Theory and Practice Udi Helman, Benjamin F. Hobbs, and Richard P. O’Neill

27

65

95

139 141

179

6.

The Cost of Anarchy in Self-Commitment-Based Electricity Markets Ramteen Sioshansi, Shmuel Oren, and Richard O’Neill

245

7.

Market Power and Market Monitoring Parviz Adib and David Hurlbut

267

8.

Demand Participation in Restructured Markets Jay Zarnikau

297 v

Contents

vi

PART III: Capacity, Resource Adequacy and Investment 9.

325

Resource Adequacy: Alternate Perspectives and Divergent Paths Parviz Adib, Eric Schubert, and Shmuel Oren

327

10.

The Evolution of PJM’s Capacity Market Joseph E. Bowring

363

11.

Resource Adequacy and Efﬁcient Infrastructure Investment Alan Moran and Ben Skinner

387

417

PART IV: Market Design Issues 12.

Promoting Electricity from Renewable Energy Sources – Lessons Learned from the EU, United States, and Japan Reinhard Haas, Niels I. Meyer, Anne Held, Dominique Finon, Arturo Lorenzoni, Ryan Wiser, and Ken-Ichiro Nishio

419

13.

Distributed Generation and the Regulation of Electricity Networks Dierk Bauknecht and Gert Brunekreeft

14.

Global Climate Change and the Electric Power Industry Andrew Ford

499

15.

Reform of the Reforms in Brazil: Problems and Solutions João Lizardo R. Hermes de Araújo, Agnes Maria de Aragão da Costa, Tiago Correia, and Elbia Melo

543

Index

573

EBL

469

Contributors

Parviz ADIB recently joined Automated Power Exchange (APX, Inc.) as a Director to enhance APX’s activities with regard to power and environmental markets. Previously, he was Director of Wholesale Market Oversight at the Public Utility Commission of Texas (PUCT) where he was engaged in designing competitive electricity market rules, providing policy recommendations to the Commission, and supervising wholesale market monitoring and enforcement functions. Prior to joining the Commission, he taught graduate and undergraduate courses at the University of Texas at Austin. He and two other Commission staff members were the recipient of the Center for the Advancement of Energy Markets’ Unsung Heroes Award for their leadership as a civil servant, which recognized them for making a signiﬁcant difference in gas and electric competition in 2005. Dr Adib’s main research interests include energy and public policy, efﬁcient operation of restructured electricity markets, and effective market monitoring and enforcement mechanism. He has published book chapters and papers and prepared numerous presentations and policy recommendations on these topics, notably covering Restructured Electricity Markets. Dr Adib completed his BA and MS studies in Economics at the Tehran University and holds a PhD in Economics from the University of Texas at Austin.

EBL

João Lizardo R. Hermes DE ARAÚJO is Director-General of the Centre for Electric Energy Research (CEPEL), a non-proﬁt research center mostly funded by the Eletrobrás Group, under leave from the Federal University of Rio de Janeiro (UFRJ) where he is full professor. Former appointments include Research Director at the Institute of Economics of UFRJ (IE/UFRJ), head of the Energy Group at IE/UFRJ, head of the Graduate Programme on Energy Planning at the Co-ordination of Graduate Programmes in Engineering at UFRJ (COPPE/UFRJ), head of the Systems and Computer Science Programme at COPPE/UFRJ. He has been a visitor at SPRU, LBL, IEJE (Grenoble), and Imperial College. Professor Araújo’s main research interests include infrastructure regulation, particularly electricity market reform, energy policy, technology innovation and diffusion, energyefﬁcient technologies; he has also worked on energy modeling and scenario-building, heuristic programming, and stochastic processes. Professor Araújo has numerous reports and publications on Brazilian energy policy and reform, including several books including the preceding book to this volume on Electricity Market Reform (2006). He has edited a UNESCO book on Energy Planning Methods (1984), and published in Energy Policy and Technological Forecasting and Social Change, among others. He is a senior researcher and referee for the Brazilian National Research Council, and referees several Brazilian scientiﬁc Journals. Dierk BAUKNECHT is a research fellow at the Oeko-Institut’s Energy and Climate Group in Germany. He is also a member of the research group Transformation and Innovations in Power Systems (TIPS), funded by the German Federal Research Department, and a PhD student with the Sussex Energy Group at the University of Sussex, UK. Before joining the institute he was modeling manager with a UK-based power market consultancy, vii

viii

Contributors

providing electricity wholesale price and generation models, forecasts, and market analysis for Europe’s liberalizing electricity markets. He works on power plant investment models, network regulation, decentralized power generation, innovation research, and transformation of energy systems. One of his main research questions is how decentralized generation can be integrated into networks and markets. He graduated in political science at the Free University of Berlin and holds an MSc in Science and Technology Policy from SPRU at the University of Sussex, UK. Joseph E. BOWRING is the Market Monitor for PJM, where he is responsible for all aspects of market monitoring. He is the principal author of the annual MMU State of the Market Report and regularly ﬁles reports on market issues with the Federal Energy Regulatory Commission (FERC). Prior to joining PJM, he worked as an independent consulting economist, with an emphasis on restructuring issues in the electric and gas industries. He worked as the Chief Economist for the New Jersey Department of the Public Advocate’s Division of Rate Counsel, as an economist with the New Jersey Board of Public Utilities, and as an economist at the US Department of Energy, Energy Information Administration. He taught economics as a member of the faculty at Villanova University and at Bucknell University. He received his PhD and MA degrees in Economics from the University of Massachusetts. Gert BRUNEKREEFT is Professor of Energy Economics at Jacobs University Bremen in Germany and Director of the Bremer Energie Institut. Before joining Jacobs University, he was Senior Economist for the energy company EnBW AG and held research positions in applied economics at Tilburg University, the University of Cambridge, and Freiburg University. He is Associate Researcher to a number of research centers and is Associate Editor for the Competition and Regulation in Network Industries. Professor Brunekreeft’s main research interests are in industrial economics, regulation theory, and competition policy of network industries, especially electricity and gas markets. Current issues concern the economics of vertical unbundling and the relation between regulation and investment. He is the author of several books and has published widely in academic journals, including Journal of Regulatory Economics, Utilities Policy, Oxford Review of Economic Policy, and Energy Journal. He holds a degree in economics from the University of Groningen and a PhD from Freiburg University, both in economics.

EBL

Hung-Po CHAO is Director, Market Monitoring at ISO New England. Before joining ISO New England in 2005, Dr Chao was with Electric Power Research Institute (EPRI) and a Consulting Professor at Stanford University. Dr Chao’s research interest includes economic analysis of market organizations, pricing and incentive mechanisms, risk management, energy and environmental analyses, and electricity market restructuring. He was a recipient of the Franz Edelman Management Science Achievement Award from INFORMS in 1988. He has written numerous papers on aspects of electricity markets and published widely in academic journals including Journal of Regulatory Economics, Operations Research, and American Economic Review. He is a co-author/editor of a book Designing Competitive Electricity Markets, by Kluwer Academic Publishing, in 1998.

Contributors

ix

Dr Chao holds a PhD in Operations Research and Economics (minor) from Stanford University. Nigel CORNWALL is Founder and President of Cornwall Energy Associates, a consulting and publishing business. He has extensive experience of electricity sector restructuring in the United Kingdom and internationally in both the public and private sectors. Prior to his return to the United Kingdom in the late 1990s he spent several years in Australia and New Zealand working in energy sector reform and abroad as a freelance consultant working in a range of emerging and developed markets. His current work is focused on the British market and retail market competition, and he was one of the main negotiators for suppliers during the development process of NETA. He has worked with a broad range of clients in the United Kingdom, especially regulators and smaller participants. He specializes in issues connected with market design, including the role of transmission and system operations in deregulated markets and the associated legal, regulatory, commercial, and governance framework. He has a BA in modern history from Pembroke College, Oxford, and an MSc in politics and administration from Birkbeck College, London. Tiago DE BARROS CORREIA is Senior Economist at the Brazilian Ministry of Mines and Energy, and specializes in energy trade and auction theory. His professional experience also includes working at the Interdisciplinary Nucleus for Energy Planning – NIPE of the State University of Campinas – Unicamp. His main research interests include electricity market reform and liberalization, demand and price forecasting, integrated resource planning, and energy auctions. He completed his BA studies in Economics and his MS studies in Energy Planning, both at the State University of Campinas (Unicamp).

EBL

Aad F. CORRELJÉ is an associate professor at the Economics of Infrastructures section of the Faculty Technology, Policy and Management at the TU Delft in the Netherlands. He is a research fellow with the Clingendael International Energy Programme (CIEP) of the Netherlands Institute for International Affairs, a member of the Editorial Board of Energy Policy, and an instructor at the Florence School of Regulation. Dr Correljé has been involved in academic research and teaching and in consulting and advising on many energy- and water-related issues, including the political economy of gas and power market developments, drinking water supply and sanitation, geopolitics of energy, sustainability policy, market design and regulatory issues, and ﬂooding protection policy. He has written a major study on the history of natural gas in the Netherlands and has contributed to the European Commission on energy supply security. After a Masters in International Relations, he received his PhD from the Centre for International Energy Studies (EURICES) of the Erasmus University in Rotterdam in 1994. Agnes Maria DE ARAGÃO DA COSTA is Senior Economist at the Brazilian Ministry of Mines and Energy, and specializes in economics of the energy and mining sectors. Her professional experience also includes working at a Brazilian bank with project ﬁnance in the energy sector. Her main research interests comprise electricity market reform and liberalization, environmental and energy policy, demand and price forecasting, integrated resource planning, ﬁnancing of energy projects in developing countries, and private sector development. She graduated in Economics at the Federal University of Rio de Janeiro (UFRJ) and completed her MS studies in Energy Economics at the University of São Paulo (USP).

x

Contributors

Laurens J. DE VRIES is an assistant professor at the Faculty of Technology, Policy and Management of Delft University of Technology. He researches and teaches about market design and policy issues in the electricity and gas markets. His research focuses primarily on generation adequacy and capacity mechanisms, but has been expanding to include the process of market design, the structuring of incentives for network expansion, and retail consumer behavior. Dr De Vries studied Mechanical Engineering at Delft University of Technology, received a Master of Environmental Studies Degree from the Evergreen State College (Olympia, Washington State, USA) and a PhD from Delft University of Technology. Dominique FINON is Senior Fellow in Economics at CNRS, France. Former head of the Institute of Energy Economics (CNRS) in Grenoble University from 1990 to 2002, he is presently member of the Center of Research on Environment and Development, CIRED, of the School of Social Sciences (EHESS) in Paris, and Director of the LARSEN (Laboratoire d’Analyse économique des Réseaux et Systèmes Energétiques), a joint institute of CNRS, EdF, and Paris University. His main research ﬁelds are the efﬁciency of the market reforms in the electricity and gas industries (market rules, long-term competition and investment, reliability and capacity adequacy) and the design of public policies in market environment (promotion of renewables, energy efﬁciency, CO2 ). His former researches dealt with technological policies, nuclear economics, and energy modeling. He has published numerous academic and professional papers and co-edited several books. He is the present chair of the French Association of Energy Economists. Dr Finon has a diploma of Engineering from Ecole Centrale de Lyon (1971) and a master in Economics from Lyon University (1971). He has a PhD. in Energy Economics from the Grenoble University (1976) and received Docteur d’Etat in Economics in 1988.

EBL

Andrew FORD is Professor of Environmental Science at Washington State University. His previous appointments have been in the Systems Management Department at the University of Southern California and the Energy Policy Group at the Los Alamos National Laboratory. His sabbaticals include visits to the Corporate Planning Department of the Paciﬁc Gas and Electric Company, the Energy Markets Group at the London Business School, and the Sloan School of Management at the Massachusetts Institute of Technology. Dr Ford’s main research and consulting interests involve the use of dynamic simulation modeling to aid in policy formulation in the electric power industry. He teaches modeling with an emphasis on energy and environmental problems in the West. He is the author of the Island Press text on Modeling the Environment. He uses the system dynamics approach to modeling and is the recipient of the Jay W. Forrester Award for outstanding contribution to the ﬁeld of system dynamics. Dr Ford earned his Doctorate from the Public Policy and Technology Program at Dartmouth College in 1975. Reinhard HAAS is an associate professor at the Energy Economics Group, Institute of Power Systems and Energy Economics, Vienna University of Technology. He specializes in energy policy strategies and the economics of the electric power sector. He is teaching Energy Economics, Regulation, and Competition in Energy markets, and Energy Modeling. His main research areas are evaluation of dissemination strategies for renewables, modeling paths toward sustainable energy systems, liberalization of energy markets, and energy policy strategies. He has published various papers in reviewed international journals on

Contributors

xi

these topics for the past 15 years. He has coordinated and coordinates projects for Austrian institutions as well as the European Commission and the International Energy Agency. He holds a PhD in Energy Economics from Vienna University of Technology. Anne HELD is a research associate at the Department of Energy Policy and Energy Systems at Fraunhofer Institute for Systems and Innovation Research in Karlsruhe, Germany. She joined the institute in December 2004 for realizing her master thesis about the evaluation of renewable promotion schemes in the European electricity market. Her research focuses on promotion strategies for renewable energy sources, the development of techno-economic models in the ﬁeld of renewable energies, and the use of renewable energy sources in the context of climate policy. She is currently working on various international research projects in this ﬁeld funded by the European Commission and on her PhD thesis. Anne Held holds a degree in Industrial Engineering and Management from the Technical University of Karlsruhe. She has authored several publications in the ﬁeld of renewable energy. Udi HELMAN is an economist at the FERC. He has worked extensively on US electricity market design issues, including auction markets for wholesale energy, ancillary services and installed capacity, transmission usage pricing, and market rules for transmission property rights. His major projects have included the initial design and redesign of the Independent System Operator (ISO) New England markets, FERC’s Standard Market Design initiative, the development of the Midwest ISO markets, and the design of longterm ﬁnancial transmission rights. Dr Helman has authored or co-authored a number of articles and book chapters on electricity regulatory reform. These include “Regulatory Reform of the US Wholesale Electricity Markets,” in M.K. Landy, M.A. Levin, and M. Shapiro, eds., Creating Competitive Markets: The Politics of Regulatory Reform (Brookings Institution Press, 2007); “Independent System Operators in the USA: History, Lessons Learned, and Prospects,” in F. Sioshansi and W. Pfaffenberger, eds., Electricity Market Reform: An International Perspective (Elsevier, 2006); “Market Power Monitoring and Mitigation in the US Wholesale Power Markets”, Energy (May-June 2006); and “Dispatchable Transmission in RTO Markets,” IEEE Transactions on Power Systems (February 2005).

EBL

Benjamin F. HOBBS is Professor in the Department of Geography & Environmental Engineering of the Johns Hopkins University. He also holds a joint appointment with the Department of Applied Mathematics & Statistics. Previously, he was Professor of Systems Engineering and Civil Engineering at Case Western Reserve University. He has also been a member of the research staffs of Brookhaven and Oak Ridge National Laboratories. Dr Hobbs’ research interests are in electric sector policy and planning, and ecosystem management. He is a member of the California ISO Market Surveillance Committee, Scientiﬁc Advisor to the ENC Policy Studies Unit, and member of the Public Interest Advisory Committee of the Gas Technology Institute. His recent books are Energy Decisions and the Environment, a Guide to the Use of Multicriteria Methods (with P. Meier, Kluwer, 2000) and The Next Generation of Electric Power Unit Commitment Models (edited with M. Rothkopf, R. O’Neill, and H-p. Chao, Kluwer, 2001). His PhD is in Environmental System Engineering from Cornell University. David HURLBUT is a senior analyst at the National Renewable Energy Laboratory in Golden, Colorado. Formerly, he was senior economist with the PUCT, where he worked

xii

Contributors

in the Commission’s wholesale market oversight group on market power, market design, and renewable energy. Dr Hurlbut’s main research and consulting interests include renewable energy policy, electricity market reform, market power and antitrust issues, price transparency, energy efﬁciency, state energy policy, water policy, and factors inﬂuencing electricity market transformation toward energy efﬁciency and renewable energy. He authored numerous reports for the PUCT dealing with market power and renewable energy. He also led a number of investigations and rulemaking projects at the Commission. He has a Master of Public Affairs and a PhD in Public Policy from the University of Texas at Austin, as well as a degree in journalism from the University of Houston. Arturo LORENZONI is Professor of Energy Economics at the Department of Electrical Engineering of the University of Padova, Italy. He is also Research Director at the Centre for Research on Energy and Environmental Economics and Policy of Bocconi University in Milan. His scientiﬁc interests are related to the technical, economic, and regulatory aspects of the energy sector and to its overall efﬁciency, with particular attention to the electric power system, the utilities, and the development of renewable energy sources. He is the author of many papers on the regulation, liberalization, and organization of the electricity supply industry. He has directed various research projects for DG Research and DG TREN of the European Commission. He has degrees in Electrical Engineering and in Energy Economics at the University of Padova and got a Masters in Energy and Environmental Economics at the Scuola Superiore ENI Enrico Mattei in Milan. He was Study Fellow at SPRU, University of Sussex.

EBL

Elbia MELO is a member of the Board of the Brazilian Power Exchange Chamber (Câmara de Comercialização de Energia Elétrica). She worked at Ministry of the Mines and Energy as Principal Economist (2003–06), and at the Ministry of Finance as Senior Economist (2001–03). At the Federal Electricity Regulatory Agency – ANEEL (2000–01) – she worked as Assessor in the Market and Economic Studies Department. Her main interests include electricity market reform and liberalization, demand and price forecasting, integrated resource planning, and energy auctions. She has an MA and a Bachelor in Economics and a PhD in Production Engineering from Federal University of Santa Catarina. Niels I. MEYER is Professor Emeritus of Physics at the Technical University of Denmark (TUD). He was vice-president and Dean of Natural Sciences at TUD in the 1970s and 1980s, President of the Danish Academy of Technical Sciences for 6 years in the 1970s, and has been member of the Club of Rome since 1972. From 1982 to 1993 Dr Meyer has chaired Danish government committees for renewable energy and from 1998 to 2002 he was member of the Danish Regulatory Committee for the Energy Sector. Dr Meyer’s main research has been concerned with solid-state physics in the 1960s, but since the early 1970s he switched to the ﬁeld of energy and environment, with a focus on renewable energy and planning for sustainable energy development. Dr Meyer has authored more than 80 international papers in the ﬁeld of solid-state physics and energy. In recent years his main research has been concerned with the consequences of the liberalized energy market in EU in relation to supply security, support for renewables, and global warming. He has PhD and DSc degrees from the Technical University of Denmark.

Contributors

xiii

Alan MORAN is the Director of the Deregulation Unit at the Australian policy review organization, the Institute of Public Affairs (IPA). Since 1996, Dr Moran has also been the Convenor of the IPA’s Energy Forum and has published studies into many different aspects of energy regulation. Among these is the Australian chapter to a book on world electricity markets, Electricity Market Reform edited by Sioshansi and Pfaffanberger. He has also published widely on other regulatory issues, including telecommunications and housing regulation and his book on the latter, The Tragedy of Planning, was launched by the Federal Treasurer in August 2006. Before his present role, he was a senior ofﬁcial in Australia’s Productivity Commission and Director of the federal Ofﬁce of Regulation Review. Subsequently, he played a leading role in the development of energy policy and competition policy review as the Deputy Secretary (Energy) in the Victorian Government. He was educated in the United Kingdom and has a PhD in transport economics from the University of Liverpool and degrees from the University of Salford and the London School of Economics. Ken-ichiro NISHIO is a scientist at the Socio-economic Research Center of the Central Research Institute of Electric Power Industry (CRIEPI), a research organization based in Tokyo, Japan, specializing in energy technologies, economics, and policy. He has also conducted research as a visiting scholar at the Lawrence Berkeley National Laboratory (LBNL) in California, USA. Mr Nishio’s main research interests include resource assessment of renewable energy, renewable energy policy, energy demand, end-use technologies such as heat pump and distributed generations, and more generally modeling of energy systems. He has authored reports on analyses of renewable energy development under the renewable energy policy in Japan, and energy conservation potential by introducing energy-efﬁcient technologies in the residential sector. He has a Masters degree in Electrical Engineering from the University of Tokyo.

EBL

Richard P. O’NEILL is Chief Economic Advisor at the FERC. His previous positions included Chief Economist and Director of the Ofﬁce of Economic Policy, Director of the Commission’s Ofﬁce of Pipeline and Producer Regulation. Prior to joining FERC he directed oil and gas analysis, including the development of software systems, oil and gas resource analysis, energy modeling systems, analysis of natural gas markets, and oil and gas forecasting at the Energy Information Administration. He has held a number of positions including teaching and work with the World Bank His work has focused on open access, restructuring, competition, performance-based incentive regulation, and market design. His published work has appeared in academic and professional journals and books in the areas of Applied Mathematics, Optimization, Operations Research, Management Science, Computer Science, Energy, Electrical Engineering, Economics, and Law. He has a BS in chemical engineering, an MBA, and a Doctorate in operations research with minors in mathematics, statistics, economics, and accounting, all from the University of Maryland. Shmuel OREN is the Earl J. Isaac Chair Professor in the Science and Analysis of Decision Making in the Industrial Engineering and Operations Research department at the University of California, Berkeley. He is the Berkeley site director of PSERC – a multi-university Power System Engineering Research Center sponsored by the National Science Foundation and industry members.

xiv

Contributors

He has published numerous articles on aspects of electricity market design, planning, and regulation and has been a consultant to various private and government organizations including the Brazilian Electricity Regulatory Agency (ANEEL), the Alberta Energy Utility Board, the Polish system operator (PSE), and the electric power research institute (EPRI). He currently serves as Senior Adviser to the Market Oversight Division of the PUCT, and a consultant to the Energy Division of the California Public Utility Commission (CPUC). He holds BSc and MSc degrees in Mechanical Engineering from the Technion in Israel and also an MS and PhD degrees in Engineering Economic Systems from Stanford University. He is a fellow of the Institute of Electrical and Electronic Engineers (IEEE) and of the Institute for Operations Research and Management Science (INFORMS). Wolfgang PFAFFENBERGER is retired professor of economics at university Oldenburg Germany and adjunct professor at Jacobs University Bremen for Economics (European Utility Management). His main interests include economics of competition in network industries and the electricity and natural gas industries in particular. He has published a textbook on electricity economics and co-authored a textbook on energy economics (both in German) and has written numerous studies and articles on macroeconomic implications of energy and environmental policy, on various aspects of market opening in the electricity supply industry (ESI), and on the special problems of the ESI in transformation countries. Dr Pfaffenberger holds a Diploma in Economics from Freie Universität Berlin where he also got his PhD. He is Honorary Doctor of the Faculty of Economics of the State University of Novosibirsk in Russia.

EBL

Michael POLLITT is University Reader in Business Economics at the Judge Business School, University of Cambridge, and a Fellow of Sidney Sussex College, Cambridge. Dr Pollitt is specialist in utility regulation and has written extensively on the regulation of electricity networks and the lessons from the deregulation of the electricity sector. He is currently an external economic advisor to the UK energy regulator, Ofgem, and Assistant Director of the ESRC Electricity Policy Research Group. He is the author of Ownership and Performance in Electric Utilities (OUP, 1995), a co-author of A Single European Electricity Market? (CEPR, 1999), and co-editor of Future Electricity Technologies and Systems (CUP, 2006) and Delivering a low carbon electricity system (CUP, 2008). He has advised the World Bank, leading energy companies, and several national regulators on energy regulation and reform. In 2005 and 2006 he served as Acting Executive Director of the ESRC Electricity Policy Research Group. He holds a DPhil in Economics from the University of Oxford. Eric SCHUBERT is Director, Regulatory Affairs, ERCOT, for BP Energy Company. His professional experience includes working at the PUCT, Research and Planning Consultants, the Chicago Board of Trade, and Bankers Trust Company. During his tenure at the PUCT, Dr Schubert was project leader in Commission proceedings involving wholesale market design and resource adequacy issues in ERCOT. In recent years, Dr Schubert has authored articles on wholesale market design and resource adequacy issues with a focus on the evolution of electricity deregulation in a policy setting. Earlier in his career, Dr Schubert published articles on ﬁnancial history with a focus on the Mississippi and South Sea Bubbles and the evolution of international ﬁnancial markets in Western Europe in the eighteenth Century. He has a PhD in Economics from the University of Illinois at Urbana-Champaign.

Contributors

xv

Harry SINGH is Senior Advisor at the FERC where his responsibilities include oversight of electric markets, providing expertise on issues related to market power and manipulation, and advising the Commission on policy issues in electric markets. Dr Singh has worked on implementing several initiatives related to the Energy Policy Act of 2005. These include rulemaking to implement long-term FTRs in RTO markets and an inter-agency report to Congress assessing the competitiveness of electric markets. Harry has also been involved in making recommendations on various investigations; drawing distinctions between manipulation, market power, and market design ﬂaws. He contributes regularly to the Commission’s policies dealing with California. Prior to joining FERC, he held various positions at PG&E Corp and its subsidiaries. His assignments included work with different business units at PG&E, the regulated utility. Harry was a key member of the start-up team that implemented the California ISO and Power Exchange. His subsequent assignments were at PG&E Energy Services and PG&E National Energy Group including work with electric markets across the United States. He has authored or co-authored several technical papers as well as expert testimony. He holds a PhD in Electrical Engineering from the University of Wisconsin–Madison. Fereidoon P. SIOSHANSI is President of Menlo Energy Economics, a consulting ﬁrm serving the electric power sector. His professional experience includes working at Southern California Edison Company (SCE), the Electric Power Research Institute (EPRI), National Economic Research Associates (NERA), and most recently, Global Energy Decisions (GED). His main interests include electricity market reform, global climate change, demand and price forecasting, integrated resource planning, energy efﬁciency, and renewable energy technologies. He co-edited with Professor Wolfgang Pfaffenberger Electricity Market Reform: An International Perspective published by Elsevier in 2006. He is the editor and publisher of EEnergy Informer, a monthly newsletter covering electric power sector, and is on the Editorial Advisory Board of The Electricity Journal where he is featured in “Electricity Currents” section. A frequent contributor to Energy Policy, he serves on the editorial board of Utilities Policy. He has degrees in Engineering and Economics, including an MS and PhD in Economics from the Krannert Graduate School of Management, Purdue University.

EBL

Ramteen M. SIOSHANSI is a postdoctoral fellow at the National Renewable Energy Laboratory (NREL) in Golden, Colorado. He has worked as a consultant for Paciﬁc Gas & Electric Company, FERC, General Electric and Global Energy Decisions, developing optimization and decision support models for operations in competitive electricity markets. He has degrees in mathematics, economics, and operations research, including an MS and PhD in operations research, from the London School of Economics and the University of California, Berkeley. Ben SKINNER is Regulatory Manager of Wholesale Markets for TRUenergy, a CLP subsidiary based in Melbourne, Australia. His professional experience includes working with the market/system operator and managing a spot market trading desk for a portfolio of generators. Ben’s main areas of regulatory interest involve the operation of the Australian National Electricity Market (NEM) and its interface with generation and transmission investment. He was involved in some of the early design groups, which developed the NEM energyonly, real-time spot market and continues to represent TRUenergy and several industry organizations in further development. He has written and presented numerous positions

xvi

Contributors

on these matters. He has visited and contrasted the NEM against the UK-NETA, ERCOT, and emerging Chinese Markets. He has a degree in Engineering and a diploma in Applied Finance. Robert WILSON retired after 40 years as a professor at the Stanford Business School. His consulting practice included projects at the Electric Power Research Institute with Drs Hung-po Chao and Shmuel Oren, with whom he co-authored the chapter in this volume. Dr Wilson’s academic research focuses on game theory. His consulting addresses issues of market design, including several projects in the power industry. Professor Wilson has authored numerous articles, and a book on Nonlinear Pricing (Oxford Press, 1993). He received a DBA degree from Harvard and honorary degrees from Bergen and Chicago. He was elected to membership in the National Academy of Sciences and designated Distinguished Fellow of the American Economic Association. Ryan WISER is a scientist at Lawrence Berkeley National Laboratory. Prior to his employment at Berkeley Lab, Dr Wiser worked for Hansen, McOuat, and Hamrin, Inc., the Bechtel Corporation, and the AES Corporation. Dr Wiser leads research in the planning, design, and evaluation of renewable energy policies, and on the costs, beneﬁts, and market potential of renewable electricity sources. His recent analytic work has included studies on the economics of wind power; the treatment of renewable energy in integrated resource planning; the cost of state-level renewables portfolio standards; trends in solar costs in California; and the risk mitigation value of renewable electricity. Dr Wiser regularly advises and consults with state and federal agencies in the design and evaluation of renewable energy policies, and is an advisor to the Energy Foundation’s China Sustainable Energy Program. Dr Wiser has published over 200 journal articles and research reports. His work has been quoted in the Wall Street Journal, USA Today, Washington Post, New York Times, LA Times, and numerous other publications. He received a BS in Civil Engineering from Stanford University and holds an MS and PhD in Energy and Resources from the University of California, Berkeley.

EBL

Jay ZARNIKAU is President of Frontier Associates, a consulting ﬁrm serving utilities, retail electric providers, large industrial energy consumers, and retail trade associations in the design and evaluation of energy efﬁciency programs, retail market strategies, electricity pricing, demand forecasting, and energy policy. He also teaches applied statistics at the University of Texas at Austin as a part-time Visiting Professor. Dr Zarnikau formerly served as a vice-president at Planergy, as a manager at the University of Texas at Austin Center for Energy Studies, and as a division director at the PUCT. His publications include articles in The Energy Journal, Resource and Energy Economics, Energy Economics, IEEE Transactions on Power Systems, Energy Policy, Energy, and The Electricity Journal. He assisted in creating ERCOT’s Demand Side Working Group, and served as its ﬁrst co-chair. He has a PhD degree in Economics from the University of Texas at Austin.

Foreword: Liberalization and Regulation in Electricity Systems – How can We get the Balance Right? MICHAEL POLLITT Judge Business School and ESRC Electricity Policy Research Group, University of Cambridge

I am very grateful to Fereidoon (Perry) Sioshansi for asking me to write a foreword to this excellent volume. Perry has done scholars of electricity reform a great service by drawing together another excellent collection of chapters on various aspects of reform in electricity markets across the world. Electricity liberalization continues to be one of the longest running and most interesting set of multi-country micro-economic experiments. While most of these national experiments are on-going, some are mature enough to no longer be considered experiments, and many have been running long enough to give rise to preliminary results. Economic analysis is well served by well-informed and detailed analyses of these experiments, such as appear in the pages of this book. In this brief foreword, I want to explore the issue of the balance between liberalisation and regulation in electricity systems, which is the essence of much of the detailed policies which are implemented in the sector (Pfaffenberger also raises the issue of this balance in his preface to this volume). By liberalisation I take to mean the use of market or quasi-market mechanisms as part of a reform of the sector, by regulation I take to mean intervention to restrain the operation of market prices or to set standards (e.g. for quality or system security) at variance with those that would otherwise have operated in the absence of regulation. I use the word “system” rather than “market” to indicate the extent of the system covered by a single regulator or system operator, which may or may not involve a market. Thus national or regional electricity systems and ISO areas would be included in what I mean by an electricity system. In looking at the issue of balance, let me say where I am coming from on this: First, liberalisation of electricity systems typically happens within a context of regulation. There is no such thing as complete deregulation of electricity markets, at least for systems of any signiﬁcant size. Even among the leaders in global electricity liberalisation – UK, Texas or Norway – what we observe is a liberalisation process which has gone further than in many other jurisdictions but what this means is that the balance between liberalisation and regulation has been shifted further in favour of liberalisation than elsewhere. Truly private networks (e.g., in Woking in the UK) or unregulated networks (e.g. in Mogadishu, within a failed state) can be observed but most systems are bigger than these. Second, the extent of liberalisation of electricity systems is a choice variable, generally bounded by what might be possible. As has been amply demonstrated by the leading countries in electricity reform, there seems to be a lot of freedom to choose how much liberalisation to introduce. European countries in particular, at least initially, exhibited

EBL

xvii

xviii

Competitive Electricity Markets

the full range of responses from virtually no change from a state vertically integrated monopoly to the creation of some of the most liberalized markets in the world. Among developing countries, Argentina and Chile exhibited substantial reform at an early stage while many more developed and developing countries have made less progress. Third, while liberalisation is a choice variable it is a variable conditioned by national institutional factors. Correlje and De Vries discuss this more fully in their chapter. Institutional factors do appear to constrain jurisdictions in their reform schemes. In general, only jurisdictions with substantial pro-competitive histories have attempted the most extensive set of reform measures. Arguably only Latin America provides any examples of extensive electricity reform within a tradition of state intervention – with Argentina being the most interesting example of this. I will explore the extent to which reform choices may be constrained by wider institutional factors later in this chapter. Fourth, it is difﬁcult for me not to write from a UK perspective. Although I am familiar with the reform experience in several jurisdictions, I naturally judge what might be possible on the basis of the experience which I have observed most closely. The advantage of doing this is that that UK has an honourable tradition of leading in the area of energy market reform, the disadvantage is obvious. Increasingly I am interested in thinking about why it is that so few of the jurisdictions that have attempted to replicate the UK experience have been successful. The explanations that naturally suggest themselves include (1) that reforms have not been extensive enough or (2) that a reform package that might be effective in the UK gives rise to perverse results elsewhere. Several US markets would seem to be examples of the former, whereas developing countries might suffer from the latter problem. A ﬁnal more general explanation (3) might be that institutional constraints – particularly with respect to the initial ownership structure of the industry – mean that while UK style reforms are technically possible, they would not be implemented because of the political economy of the interest groups that exist in the countries concerned. Several major European countries, such as France, Spain and Germany would seem to in this category. For example, in France the starting point is a powerful and well regarded state owned company, while in Germany the starting point involves economically powerful private companies. I will continue by examining the case for liberalisation of electricity systems, then look at the case for regulation, offer a brief evaluation of what the evidence on where the balance might be drawn, before considering what the future might hold for electricity liberalisation.

EBL

The Case for Liberalisation As Chao, Oren and Wilson remind us in their chapter, the past for most electricity systems involved a high degree of vertical integration – particularly between generation and transmission, but also between distribution and retail and frequently between all four stages of production. The economic arguments for large vertically integrated electricity companies, which were signiﬁcant in size within political jurisdictions rested on a claim that vertical economies were signiﬁcant. Many studies examined the nature of these vertical economies and most, if not all, concluded that vertical economies between generation and transmission were signiﬁcant (e.g. Kaserman and Mayo, 1991). This provided arguments for the integration of what we now know to be potentially competitive generation and monopoly transmission networks. Papers which claim vertical economies continue to be written, e.g. on the regional Japanese electric power companies where a very high degree of vertical integration continues (see Nemoto and Goto, 2004).

Foreword: Liberalization and Regulation in Electricity Systems

xix

Experience with liberalisation shows us that there were indeed costs to vertical separation, both in generation and transmission and between distribution and retail (see Newbery and Pollitt, 1997 and Domah and Pollitt, 2001). However these costs had to be incurred to achieve the beneﬁts of more competition in generation and also more competition in retail. Studies showing vertical economies between the various stages of electricity supply are mostly not capable of modelling these beneﬁts properly, and indeed in the case of studies where there are no non-integrated companies with whom performance of integrated utilities can be compared are seriously mis-speciﬁed. The case for competition in generation rests on the beneﬁts of competitive markets generally. More precisely, a central claim is that competition drives efﬁciency gains, which can be substantial given that generation can be as much as 65% of value added in the sector. Efﬁciency gains are of two main types: cost savings arising from the more efﬁcient operation of existing assets and those arising from the choice of cheaper technologies for new generation. Both can be signiﬁcant, even more so where initial operation has been in the hands of inefﬁcient publicly owned utilities. Heavily regulated integrated utilities are subject to a potential gold plating problem if they are privately owned (following Averch and Johnson, 1962) and subject to the pressure to be instruments of government industrial policy which supports expensive home grown contractors and designs, rather than generic scalable technologies which are available off the shelf (see Henderson, 1977 and Green, 1995 on the highly expensive British nuclear programmes). Clearly the uptake of CCGT in competitive generation has been a great success in liberalized electricity markets, as has the cutting of support for expensive untried clean coal technologies (which would not have been as clean as CCGT) and the curtailing of the uneconomic roll out of nuclear power (Newbery and Pollitt, 1997). Competitive generation has clearly revealed the price of different technologies and caused clear choices to be made in line with market principles. Electricity economists are keen to point out how different electricity is to other commodities and hence what some of the problems of letting the market operate might be. Indeed Chao, Oren and Wilson note that the regulatory compact in the US electricity sector was that monopoly operation of electricity networks was essentially the price that was paid by regulators in return for smoothing of retail electricity prices. They suggest that electricity prices would be much more volatile in the absence of regulation and that this would be politically unacceptable. This is undoubtedly a powerful view of competition in retail electricity markets, which resonates today, even within the most fully liberalized markets. However it is a rather quaint given rising incomes and the increasing availability of sophisticated ﬁnancial instruments. It may also be a view arising from US states where low power prices have been traditionally based on access to cheap coal (Joskow, 1997). Consumers in these states seem to have little to gain by deregulation and market extension to high price states. By contrast the enthusiastic advancement of market liberalisation in California was driven by high electricity prices in a state where demand was growing rapidly and there no cheap resources for power generation. Ambivalent views towards electricity reform are rather at odds with economists’ normal enthusiasm for market-determined prices. What full retail competition (i.e. including households) in both the UK, Nordic countries, New Zealand and Texas reveals is that consumers can be content to pay “volatile” electricity prices as long as there is a perceived beneﬁt from retail competition. Undoubtedly the Chao et al. view does restrain regulators enthusiasm for proper retail competition in many US states and elsewhere. However it is based on the view that competition is ok in most markets, but somehow not ok in retail energy markets. This is rather odd, as economists (though not politicians) are happy to see volatile gasoline prices, but not volatile residential electricity (or natural gas?) prices. It is difﬁcult

EBL

xx

Competitive Electricity Markets

to see how retail competition at the household level can make much progress in many jurisdictions unless this view faces serious challenge. At this point it is important to be clear about what full retail competition looks like in electricity markets. This is because what passes for retail competition in many US states is such a pale version of competition as to invite ridicule. Retail competition should involve the ability to switch electricity supplier. Effective retail competition would involve the existence of, say, ﬁve or more energy companies offering retail tariffs. If a customer switches from the incumbent supplier they would receive a bill from a different company who would purchase monopoly transmission and distribution services from incumbents but be responsible for billing, contract terms, bundling of other services and the purchase of wholesale power. In many US states this is not what is meant by “retail choice” at the household level. Retail choice simply involves buying the wholesale power component of one’s energy bill from a non-incumbent. This eliminates competition in billing, bundling of other services and contract terms. This is an unusual type of competition, which is not the same as we observe in genuinely competitive markets. To make matters worse, US retail choice often involves the offering of default regulated tariffs to retail consumers which restrict switching by being regulated at a low level or having the property that if one switches, one cannot return to the default service tariff. The lesson from deregulated markets where there has been signiﬁcant customer switching is that the existence of a low regulated retail tariff discourages switching i.e., there should be a removal of price regulation of retail tariffs. The beneﬁts of household retail competition continue to be debated (Littlechild (2002) expresses the arguments for full retail competition, while Green and McDaniel (1998) present a sceptical view). Retail competition has now been revealed to be important for competition in generation. A major observation of electricity reforms is that monopoly networks are different businesses from the competitive businesses – generation and retail. The former are, when regulated effectively, low risk infrastructure businesses, while the latter are higher risk businesses subject to the normal bankruptcy risk faced by companies in competitive markets. While the vertical economies between generation and transmission are not sufﬁcient to offset the beneﬁts of competition in wholesale power markets, they do appear to be signiﬁcant between generation and retail. In particular the risk management advantages of generation and retail integration are very important, such that stand-alone retail electricity companies have struggled. Across the world large stand-alone start-up retail companies have, after some initial successes, have generally failed, as exempliﬁed by the experience of the UK, Netherlands and New Zealand. A notable retail-only business model, which has succeeded is that of the former incumbent gas company in the UK (Centrica) which has very successfully diversiﬁed into electricity retailing in the UK and in the rest of Europe. However this company has been born of an unusually competitive and unbundled gas industry in the UK. The good news is that competition in retail markets is possible; the bad news is that it is only likely to be as extensive as competition in generation because non-integrated retail companies have little chance of success at any reasonable scale. Economists are fond of pointing out the many other ways in which the detailed operation of the electricity market needs to be regulated. Thus there are issues to do with the efﬁcient operation of the transmission system and the allocation of transmission capacity and the problems of incentivising enough generation capacity to be available to limit price spikes at peak times (see Singh, this volume). The liberalisation process has revealed that market and quasi-market mechanisms can address market and regulatory failures in these areas. PJM and some other jurisdictions, for example, have successfully implemented nodal pricing

EBL

Foreword: Liberalization and Regulation in Electricity Systems

xxi

arrangements for the allocation of transmission capacity and many other markets have used price signals to make more efﬁcient use of scarce transmission capacity. In terms of the mitigation of price spikes, special mechanisms have been introduced to pay generators to make extra capacity available, either more generally via a capacity payment (e.g. in the North East US), or speciﬁcally at the system peak (e.g. in the UK). While the long-term incentive properties of some of these mechanisms are dubious, they can mitigate the short-term capacity shortage problems (this seems to be the case in New England) and provide comfort to regulators and politicians that something is being done to prevent the lights from going out. This view is supported by Adib, Schubert and Oren, in their chapter, who suggest that “there is no evidence” that capacity support mechanisms actually promote investment in the long term. Moran and Skinner’s chapter on Australia, further suggests that a market without capacity payments can work well, even though prices do occasionally spike. However as long as these spike reduction measures are not too expensive, they may be a price worth paying for the continuation of electricity reforms elsewhere in the system. Indeed, Bowring’s chapter on PJM’s capacity markets suggests that they may have a limited role to play in reassuring stakeholders that adequate capacity will be available. Bowring concludes that capacity markets are “not a panacea” but can play “a critical but circumscribed role in wholesale power market design” [my italics]. Generation and retail electricity markets yield major market advantages over vertical integration. First, they allow the efﬁcient handling of business risk. There are substantial uncertainties in the short, medium and long term in power markets. Markets are good at handling these types of risk. Indeed given the capacity of the oil market to handle much more signiﬁcant risks, it would be odd if we did not leave these to the market in electricity. Second, price has a signiﬁcant role to play in ensuring security of electricity supply. The old vertical integrated system did provide security of supply – at a cost. Most markets allow a signiﬁcant role for the price in ensuring security of supply. Around the market price insurance can be offered to those who want it, while those who want to self-insure can opt out. As the UK (2004–06), New Zealand (2001), Nordic countries (2002–03) and Chile (1998–99) have all demonstrated in recent years, retail electricity customers are willing to be exposed to signiﬁcant price volatility and play a signiﬁcant role in demand management. Third, full retail competition offers a signiﬁcant political advantage over vertical integration: the “privatisation” of the ﬁnal price of electricity. Where retail prices continue to be regulated, as they are in the absence of full retail competition or the continuation of a default service tariff, political or regulatory interference is much more likely. This is important at times of rising energy commodity prices, when the pressure will be to slow the translation of these into higher retail prices. Between 2004 and 2006, wholesale gas prices quadrupled in the UK and retail electricity and gas prices rose signiﬁcantly. However the most common response to this sharp rise was for politicians to point out that most electricity and gas consumers could cut their energy bills by switching to a cheaper energy supplier. In the absence of retail competition the pressure for politicians’ intervention would be much stronger than this. Continuation of regulation of the ﬁnal price of electricity is usually a sign of insufﬁcient competition in the wholesale power market. Removing such regulation focuses regulators attention on making the power market competitive. This is clearly demonstrated by the EU Energy Sector Inquiry, where advent of full retail competition at the household level has led the European Commission to focus on the removal of barriers to competition in wholesale power markets. Fourth, competition promotes innovation, some of which may be unexpected. Retail competition in electricity markets has promoted joint marketing of electricity and gas

EBL

Competitive Electricity Markets

xxii

where these were previously often provided by different local monopolies, the use of internet-based switching sites, payment by direct debit and the vertical integration of generation and retail (in jurisdictions where this did not previously exist). We have already noted the effect that competition in generation has had on the adoption of generation technologies. However it has also led to the refurbishment of assets, rather than their replacement (or originally scheduled closure). For example, the widespread life extension of the US nuclear ﬂeet (where no plants have shut down since 1998 and where the majority seem set to be life-extended to 2035 and beyond) is something that is clearly incentivised by competition in generation (see Joskow, 2006a). The Case for Regulation Even in the leading utility reform sector – telecoms – regulators have been slow to remove regulated retail and access tariffs to networks for competitive telecom service providers. Thus we might expect there to continue to be a strong pressure for continued regulation in the electricity sector. Demsetz (1968) asked the question: why regulate utilities? He pointed out that while natural monopoly might still exist, it could be restrained by franchising or competition for the ﬁeld in utility sectors, such as electricity. In New Zealand, as Bertram (2006) discussed in this book’s predecessor volume, the government conducted a rather interesting economic experiment. They left the regulation of the operation of the electricity distribution businesses to general competition policy. This resulted in the lines businesses signiﬁcantly increasing their charges relative to costs and absorbing all of the beneﬁts of competition in generation and retail (see Fig. 1). This experiment graphically demonstrates the fact that

EBL

16 14 12

12.90

2.11

12.34

13.17

2.25

13.68

13.63

14.10

14.17

14.11

2.40

2.42

3.10

3.13

3.10

1.89

10 4.05

3.84

4.62

4.92

5.01

4.86

4.89

4.90

2.18

2.06

1.90

1.86

1.85

1.80

1.80

1.79

4.56

4.55

4.40

4.50

4.35

4.34

4.35

4.32

Apr-94

Apr-95

Apr-96

Apr-97

Apr-98

Nov-99

Feb-00

8 6 4 2 0

Generation

Transmission

Apr-99 Line

Retail

Fig. 1. New Zealand residential electricity price c/kWh (2000 NZ$). Source: Ministry of Economic Development (2000).

Foreword: Liberalization and Regulation in Electricity Systems

xxiii

electricity networks do have signiﬁcant market power and need to be formally price regulated. It also provides a good argument for an effective sector speciﬁc regulator formally charged with implementing a price review process in networks, even in jurisdictions with competent general competition authorities. Regulation of the average prices or revenue of natural monopoly networks is not enough to ensure that network owners do not abuse their monopoly power in situations where there continues to be vertical integration of electricity networks with either generation or retail businesses. As the experience with telecoms deregulation shows, there is a need for the regulation of access to monopoly networks for generators trying to reach customers and retailers trying to source power (see Bergman et al., 1998 on telecoms). Such access pricing needs to ensure that the incumbent network asset owners can secure a return to their investment and that switching to non-incumbents occurs simply on the basis of differences in costs, prices and service offerings in the competitive parts of the supply chain. In the case of telecoms, the efﬁcient component pricing rule (Baumol et al., 1997) attempted to limit inefﬁcient bypass of the incumbent’s assets in the competitive segment in conditions where the ﬁnal price was regulated. In electricity, the problem of access pricing involves clear non-discrimination in the allocation of network access and the regulation of tariffs to reﬂect the true costs of the monopoly network. In general, energy regulators have been able to do this where their powers have been sufﬁcient, though in some continental European countries, such as Germany, there have been substantial problems reported by non-incumbent generators trying to get access terms in the early years of restructuring (see Bergman et al., 1999). In retail, there have been examples of incumbent retailers integrated with distribution wires attempting to allocate an overly generous part of the initially shared assets to the monopoly distribution wires business and hence raising access charges for non-incumbent retailers. This effectively means that the retail customers who switch to non-incumbents subsidise the retail part of an incumbent’s business. The potential size of this cost allocation problem is substantial. The UK electricity regulator, Ofgem, reallocated an amount equivalent to around 18% (of which a third was metering costs) of the controllable cost of regulated distribution from distribution wires to retail in 2000 (see Ofgem, 1999, p.17). This reﬂected the misallocation of assets and costs within incumbent integrated distribution and retail companies at the time the market was opened to full retail competition. The appropriate regulation of access charges and terms continues to drive the debate over unbundling of network assets in Europe. Successive European electricity directives in 1996 and 2003 have mandated increasingly tough unbundling requirements between the network and competitive businesses within integrated companies. The recent EU Energy Sector Enquiry highlighted access problems as a major barrier to the advancement of competition in European electricity and gas markets (European Commission, 2007). In mid-2007, the European Commission is pressing for full ownership unbundling of gas and electricity transmission networks from the rest of the system. This has been proposed to ﬁnally remove the incentive on transmission owners to favour their own generation or to not propose transmission investments, which would bring system beneﬁts at the expense of corporate proﬁts in generation. It remains to be seen whether the Commission will eventually force ownership unbundling of just system operation along an ISO model (such as PJM) or will also achieve the ownership unbundling of transmission assets to create independent TSOs (such as in the UK or under the stalled RTO model in the US). The weaker ISO model option may prevail as a compromise with countries for whom ownership unbundling of transmission assets is politically unacceptable. ISOs without

EBL

xxiv

Competitive Electricity Markets

transmission assets may do well on non-discrimination but do not solve the problem of under-investment in transmission by companies with market power in generation. Singh discusses the problems in US transmission in his chapter, putting emphasis on the need to improve transmission investment incentives and the fact that the debate in the US has begun to focus on the case for independent transmission companies as opposed to ISOs. Regulation is also important in a number of non-price areas: for example, standards for quality of service, reliability and network losses. Unless regulators enforce some sort of penalty price for the non-delivery of minimum standards, there may be incentives for services to deteriorate or not optimally improve the quality of the operation of the network. A good example of this follows. Distribution losses are paid for by retail customers who must buy extra power to cover network losses. Network losses are assumed, for charging purposes, to be uniform and constant in most distribution networks. Hence there is little advantage for competitive suppliers seeking to minimise these, resulting in market failure. In the UK, network losses in distribution did not change much over the ﬁrst ten years following liberalisation, until the regulator introduced a tougher ﬁnancial penalty for distribution losses payable by network owners (see Jamasb and Pollitt, 2007). This led to a sharp and immediate cut in distribution losses as shown in Fig. 2. The creation of competitive wholesale markets has been associated with the need to create appropriate transmission access regimes. However the rise of distributed generation (embedded within the distribution network), in response to technological change and climate change concerns, has necessitated a more considered regulatory approach to the regulation of distribution access. This is pressurising regulators to develop more ﬂexible access terms for distributed generation within the distribution network, which reﬂect its true value to the system. The regulatory issues raised by distributed generation are discussed in the chapter by Bauknecht and Brunekreeft. It is important to note that there has been a signiﬁcant improvement in the sophistication of regulation of natural monopolies since electricity reform began. The Averch-Johnson effect, the theory of regulatory capture (Stigler, 1971) and distortions of regulated price differentials (Peltzman, 1976), not to mention the rather unimpressive early attempts at limited (and hence distortionary) incentive regulation (Berg and Jeong, 1991) are lessons

EBL

8.0% 7.5% 7.0% 6.5% 6.0% 5.5% 5.0% 1990- 1991- 1992- 1993- 1994- 1995- 1996- 1997- 1998- 1999- 2000- 2001- 2002- 200391 92 93 94 95 96 97 98 99 00 01 02 03 04 Fig. 2. Distribution losses in the UK as percentage of energy delivered. Source: Jamasb and Pollitt (2007).

Foreword: Liberalization and Regulation in Electricity Systems

xxv

that now seem to be well learned by the best regulatory agencies operating in competitive markets. Indeed there does seem to be a rather clear positive correlation between the degree of development of liberalisation and the quality of the regulatory agency (see Green et al, 2005). Particular progress has been made in the incentive regulation of electricity networks, where sophisticated analysis has been combined with the use of ﬁxed term price controls, which incentivise cost reduction (Jamasb and Pollitt, 2001 and Joskow, 2005). Indeed an obvious potential gain from unbundling electricity networks from the rest of the supply chain is the facilitation of superior regulation of network businesses. This can lead to signiﬁcant cuts in regulated tariffs and/or increases in investment, as exempliﬁed by the experience of the UK, Nordic countries, Chile and New South Wales. Effective regulation not only brings direct beneﬁts through more efﬁcient networks but also through the facilitation of wholesale and retail competition across these networks. While the regulation of electricity networks is clearly necessary, electricity market liberalisation has thrown up a strong case for market monitoring (as detailed in the chapter by Adib and Hurlbut) of the operation of the competition within generation and retail markets. Market monitoring requires vigilance and quick action to correct market rules if societal welfare is to be maximised (or not signiﬁcantly reduced). Specialist energy regulators seem best able to analyse market data and to propose remedies, with detailed market rule changes sometimes delegated to system operators or wholesale market governance mechanisms if sufﬁcient regulatory threat is in place. The need for regulatory oversight of this type would seem to be more necessary in the early stages of market redesign due to the complexity and scope for gaming which exists in electricity markets. However over time regulatory oversight may be reduced as market operators become more able to self-regulate. A parallel might be drawn with ﬁnancial markets in this respect where effective self-regulation may be generally desirable within a context of appropriate legal protection and the operation of the general competition authority. Electricity markets may need to be redesigned from time to time with signiﬁcant regulatory involvement (in assessing the relative merits of different designs) as new analysis on optimal market design comes to light. An example of such a design question is raised in the chapter by Sioshansi, Oren and O’Neill who examine the relative merits of central versus self-commitment of generation units. They come down in favour of central commitment rather than self-commitment. Self-commitment was a signiﬁcant element of the New Electricity Trading Arrangements for wholesale power which replaced the Pool (begun in 1990) in England and Wales in 2001. In their chapter O’Neill and Hobbs consider the lessons from different auction markets in the US for wholesale energy and ancillary services. All such signiﬁcant redesigns need careful assessment in order to avoid imposing additional costs on electricity customers. Particular problems have been noted with the utilisation of interconnectors between jurisdictions, price signalling within power pools, withholding of capacity and mergers between energy companies. Interconnectors are a particular issue within the transmission system because they can be closely monitored. What seems to be the case, is that these can be often congested, ﬂowing in a perverse direction and not subject to appropriate capacity upgrading. In Europe, international electricity and gas interconnectors are the subject of scrutiny of regulators (see European Commission, 2007) to detect the operation of market power by integrated incumbents. Gas interconnector utilisation (or the lack of it) and non-responsiveness to large international price differentials have been particularly noted. Power pools and organised power markets can be subject to price manipulation. This is because of the repeated game nature of such interaction and the ability of ﬁrms to signal to one another within the power pool. Such accusations were regularly made of

EBL

xxvi

Competitive Electricity Markets

ﬁrms in the early days of power pool in the UK. Indeed when the regulator imposed a price cap on the incumbents in this market, they managed to hit it exactly (see Newbery, 2006). California’s ill-fated power markets are still under investigation for the alleged withholding of capacity to drive up prices (see Sweeney, 2006). Measures of the residual market power index clearly show that in many power markets incumbents have the ability to withhold enough capacity to drive up the price signiﬁcantly. Finally, power company mergers require careful assessment. This is because the sheer volume of transactions among energy companies brought about by liberalisation has the capacity to rapidly change the shape of the power sector (see Codognet et al., 2002). The increase in effective market size brought about by liberalisation in many US and European markets, requires mergers to be evaluated more carefully than would have been the case in the past. In Europe, the concentration of ownership at the European level has been the price paid to reduce concentration within individual national markets (see Jamasb and Pollitt, 2005). Meanwhile mergers between electricity and gas companies have created the potential for the sophisticated exercise of market power in the interrelation between gas and electricity markets. The highly controversial merger between EON and Ruhrgas in 2003 integrated the major supplier of wholesale gas and one of the major energy companies in Germany, eliminating a major potential competitor in the electricity market. This merger was opposed by the German competition authority but approved by the government. A similar merger between Endesa and Gas Natural in Spain in 2006 was initially approved but failed to occur. It is important to stress that while liberalisation has involved some reduction in regulation in wholesale and retail electricity, it is not necessarily the case that liberalisation and regulation are substitutes. Wholesale and retail competition may require more regulation (in terms of cost and complexity) to ensure their success. A state owned vertically integrated national monopoly may require minimal regulation, but is not consistent with a liberalised market.

EBL

Evaluating the Experience with Electricity Market Reform Globally, the evidence on the success of electricity reform is surprisingly mixed. One would have expected to ﬁnd clear evidence of reform success showing up by now in the econometric evidence. This is however difﬁcult to demonstrate as Jamasb et al. (2004) show. There are only two academic econometric studies of the price effects of electricity reform in the OECD (Steiner, 2001 and Hattori and Tsutsui, 2004). Both show some weak evidence of lower prices in the period to 1999, but it is difﬁcult to distinguish which reform steps may be contributing to this. The evidence for developing countries is even more difﬁcult. There is support for the view that incentive regulation, with privatisation and wholesale competition does raise investment, reduce losses and lower prices. However the precise interaction between reform steps is somewhat complex. The picture is complicated somewhat because pre-reform prices may have been artiﬁcially low relative to economic levels. Successful reform in these cases involves raising the price or simply holding prices constant while efﬁciency is increased. Part of the problem is that the success of electricity liberalisation might be measured on several dimensions, for example, price, quality of service, losses and investment, and may need several years of observation before a trend improvement is apparent (given “normal” year to year volatility). In recent years high commodity prices for gas further cloud the picture. Another problem is a lack of careful cost beneﬁt analyses of reforms in particular

Foreword: Liberalization and Regulation in Electricity Systems

xxvii

jurisdictions where the counterfactual has been clearly worked out and is defensible. The basic assumption that variables of interest would have been same as before in the absence of liberalisation may not be the right one. Indeed in many cases, the liberalisation may have resulted in some variables (e.g. price) rising less rapidly or indeed falling less rapidly, than would otherwise have been case. A promising way forward is to make a list of where one might think there has been a successful electricity reform. A good start would be to look at the chapters of Sioshansi and Pfaffenberger (2006). Among developing and transition countries Argentina and Chile stand out as successful poster cases. After this, the list becomes quite thin with perhaps, Peru and Columbia as additional cases. In many others there have been signiﬁcant problems (e.g. Brazil: as discussed by Araujo, Correia, Costa and Melo in their chapter), and yet others not a lot has yet been tried (e.g. Middle East, Africa, parts of India). Among developed countries, there has been a lot of reform impetus. The EU has attempted a major organised reform programme but only the UK, Norway, Sweden, Finland and the Netherlands can really be said to be making signiﬁcant progress from the pre-reform model. In most other European countries demonstrating a positive impact from electricity reforms and indeed much real enthusiasm beyond that generated by wishing to be good European citizens in complying with the Directives is difﬁcult. In North America, Texas, PJM, New England and New York demonstrate good progress in wholesale competition, but only Texas and some odd bits of Maine, Pennsylvania and Ohio have household retail choice, while many states exhibit no progress from the traditional vertically integrated model. In Maine, Pennsylvania and Ohio regulated retail tariffs continue to limit competition and “choice” is simply in the generation part of the retail bill. In Ohio one incumbent company discourages switching by highlighting consumer’s right to choose not to release data to alternative energy suppliers (on privacy grounds). Australia and New Zealand have shown enthusiasm for reform but signiﬁcant problems remain in several Australian states (e.g. Queensland, Tasmania and Western Australia where only partial liberalisation has occurred), though good retail competition is developing in Victoria and South Australia (see Moran and Skinner, this volume on Australia). In New Zealand there have been a number of legislative attempts to progress reform in the face of setbacks. In Japan progress with electricity reform continues to move very slowly. Overall the picture is one of signiﬁcant progress towards the maximum possible electricity reform being made in only a handful of jurisdictions, while reform progress has stalled at some intermediate stage in many more jurisdictions. Even some of the places where we think of the most progress as being made: parts of PJM, as well as Chile and Argentina have not got full retail competition at the household level. Correlje and De Vries suggest that, it is difﬁcult to come up with clear lessons from electricity reforms which translate globally because of what they call physical, economic and institutional factors. However there are some lessons: First, ownership unbundling of electricity transmission from the rest of the electricity network has produced clear beneﬁts in the markets where it has been tried in terms of improving access conditions for competitive generation and removing incentives to underinvest in transmission. ISOs with continued integration of transmission and generation are largely effective in improving access conditions and the operation of the transmission system but suffer signiﬁcantly from continuing under investment; Chile being an excellent example of this, among others (see Pollitt, 2004). Second, getting the market structure right in electricity generation is crucial for the success of electricity reform. This involves sufﬁcient divestiture to, say, ﬁve players by

EBL

Competitive Electricity Markets

xxviii

incumbents. Relying on new entry alone will not be enough to ensure the reduction in the market power of incumbents. Where regulators have been unable to introduce low concentration at the opening of new power markets, they have faced an uphill battle to reduce concentration. Really successful wholesale power markets either introduced competition right at the start e.g., in Argentina, PJM, Texas or undertook signiﬁcant regulatory intervention subsequently, e.g., in UK. Third, incentive regulation based on RPI-X price control of monopoly networks can deliver signiﬁcant incentives to reduce costs and can facilitate efﬁcient operation with the provision of stable cash ﬂow for new investment. Regulators that have not made best use of incentive regulation have missed out on a substantial part of the gains from ending of the vertical monopoly in electricity. This is clear from a glance at the division of the ﬁnal price of electricity between generation, transmission, distribution and retail elements in European countries. The countries with tougher incentive regulation of networks have signiﬁcantly lower network costs than those who do not (Jamasb and Pollitt, 2005). Fourth, regulation can address market failures in electricity markets such as those associated with quality of supply. The problem with regulatory incentives to meet nonprice objectives is that they may interact with one another and create perverse incentives. Regulators need to understand the power of incentives both to solve perceived problems in electricity systems and to inadvertently incentivise welfare reducing activity. Indeed part of the problem of electricity liberalisation is that it requires more sophisticated regulation to be successful. Overall the potential beneﬁts of electricity reform seem to be positive. Newbery and Pollitt (1997) conservatively estimated that generation and transmission reform in the UK reduced costs permanently by 5%. Fabrizio et al. (2007) come up with a similar ﬁgure for the reduction in generation costs due to reforms in the US. Looking at ﬁnal electricity prices in the US, Joskow (2006b) ﬁnds that competitive wholesale markets and retail competition have reduced prices (relative to their absence) signiﬁcantly, with retail competition reducing prices by 5–10% for residential cutomers and 5% for industrial customers. Focussing on just the New England wholesale power market, Barmack et al. (2007) ﬁnd a net gain of 2% of costs due to reforms. In distribution and retail, the beneﬁts are similar in order of magnitude. Domah and Pollitt (2001) identiﬁed gains of around 10% of costs for the UK. For developing countries, Toba (2007) estimated that the liberalisation of Philippine electricity generation produced a one-off gain equivalent to around 10% of GDP, while Mota (2004) estimated that the privatisation (and incentive regulation) of Brazilian electricity distribution produced a one-off gain of more than 2% of GDP. Stephen Littlechild (2000) breaks down the beneﬁt of the early years of the UK reforms in Table 1.

EBL

Table 1. Sources of price reduction to domestic users 1991/92–1998/9 Source Lower Lower Lower Lower

% generation costs distribution and transmission charges supply business margin fossil fuel levy (mainly to fund nuclear liabilities)

Total Source: Littlechild (2000).

10 9 1 9 29

Foreword: Liberalization and Regulation in Electricity Systems

xxix

Littlechild’s analysis helpfully focuses on residential customers and provides a point estimate of the impact of reform (in contrast to the NPV calculations of the Newbery and Pollitt (1997) and Domah and Pollitt (2001) studies). This indicates that the reduction in residential prices is split almost equally between lower generation costs (operating efﬁciency and fuel switching), lower distribution and transmission charges (transferred to consumers via RPI-X) and a lower fossil fuel levy, which was the subsidy to the nuclear sector to cover un-ﬁnanced decommissioning liabilities (the ending of an industrial subsidy). Thus one might say that gains were coming from competition, improved economic regulation and the elimination of inefﬁcient industrial subsidies to energy production. This highlights the fact that in many countries a clear beneﬁt of liberalisation is likely to be the end of expensive industrial policies towards the energy sector or at the very least to increase the transparency of the cost of these.

The Future of Electricity Liberalisation Electricity reform around the world has been driven by a combination of the success of the early reformers (who inspired many followers) and the extensive reform programs at the federal level in the US and at the level of the European Commission in Europe. As the difﬁculty of replicating the success of the early reformers has become more apparent, particularly in developing countries, and as the US federal program has stalled, many jurisdictions have failed to pursue electricity liberalisation to its logical conclusion. Only in the EU has the sustained commitment of the European Commission within a wider single market agenda (which encompasses all industrial sectors) and a political system where energy reforms can become part of the international bargaining process has the pressure for further electricity market reforms been sustained. However, Cornwall’s chapter highlights just how far there is to go to establish regional electricity markets (covering groups of neighbouring countries) in Europe, let alone to the Commission’s goal of a single electricity market in Europe. What is becoming increasingly clear is that electricity reform has failed to convince many of its merits (e.g. Thomas, 2006). On the basis of the evidence of the delivery of clear beneﬁts being somewhat mixed, this is hardly surprising. What seems to be the case is that the pursuit of electricity reform through to its logical conclusion is only likely to happen in jurisdictions where there is a strong ideological commitment to competition in energy markets. This will partly be driven by resource conditions – the presence of initially high costs with scope for efﬁciency gains is conducive to reform – but signiﬁcantly by whether there is a fundamental belief that electricity prices should be left to the market. This belief is partly a belief in the market itself and partly a belief in the market for energy per se. Some who believe in the market still think energy markets sufﬁciently different from others to warrant the sort of intervention that prevents the emergence of effective competition in generation and/or in retail. Often this reﬂects (or is supported by) concerns about security of supply and increasingly concerns about pollution from electricity production. What is absolutely clear is that successful through-going electricity liberalisation requires both a belief in competition and effective institutions of competition policy. Countries, such as France and Germany, will struggle to make serious progress with electricity reform unless they can change their attitude to competition in the electricity sector. This is not beyond the bounds of possibility but it will require the sort of reluctant change in attitude, which globalisation and technological change has brought about in other sectors (such as telecoms).

EBL

xxx

Competitive Electricity Markets

A new and important challenge to electricity reform is posed by climate change. As Ford discusses in his chapter, climate change is a serious issue which the power sector will be expected to address. Economically sensible policies for the internalisation of the external costs of CO2 emissions have been enacted in the EU (as the EU Emissions Trading System) and proposed in the US (various state and regional trading initiatives are progressing, with a bill for a national scheme being put before the Senate in 2003). However, with regard to the balance of liberalisation and regulation in electricity systems, climate change is a potential vehicle for the return of old-style intervention in electricity generation and in retail competition. The argument will be that the market will not invest in low carbon generation without long-term contracts for low carbon electricity generation. Such contracts would effectively eliminate competition in the wholesale power market. On the retail side, the argument will be that consumers who switch, fail to provide long-term incentives for micro-power and demand side management (DSM) investments. Regulators face a signiﬁcant challenge if they are to introduce subsidies to low carbon electricity generation which do not effectively end competition. Clearly environmental externalities should be priced properly as a ﬁrst step. Competition should be viewed as part of solution, not the problem. Encouraging price sensitivity and consumer selection between own generation, DSM measures and energy supplier would be a more effective way of mobilising extra resources (from customers) and cheaper market led responses for tackling climate change than potentially highly expensive centrally imposed investments. The potential for DSM in liberalised market environments is taken up in the chapter by Zarnikau. In the context of rising political concern about climate change, we need to heed the lessons of history on the poor track record of government backed technologies in energy. Cohen and Noll (1991) highlight the “technology pork barrel” in the US and the difﬁculty of ceasing government funding of energy technologies which fail to deliver. Fri (2003) suggests that although the theoretical case for public funding (or consumer subsidy) of energy R&D is compelling, the track record suggests that the situations where intervention is likely to have positive net present value are rather limited. He also makes the interesting observation that subsidies for strategic deployment to exploit learning economies have been of dubious economic value as learning has been just as rapid in non-subsidised “mature” technologies. Most successful innovation in electricity systems is incremental (Fri, 2003) and best left to the forces of the competition, attempts to force the pace by subsidy are likely to be expensive mistakes (especially in aggregate across the portfolio range of technological interventions). By contrast the introduction of market based incentives to abate environmental pollution in electricity have an excellent track record. The US SO2 cap and trade programme being the most notably successful of these (see Ellerman and Dubroeucq, 2004). In this context it is vital that we properly evaluate the success of the various renewable electricity support schemes implemented around the world (as for example in the chapter by Haas et al.), while continuing to press for the establishment of a sensibly high trading price for CO2. Once this price is established the requirement for large and potentially wasteful subsides to support the roll out of renewable capacity will be substantially reduced. Liberalised electricity markets have had a good run where they have been implemented. The US, UK, New Zealand, Australia and Scandinavia initially had very favourable generation capacity reserve margins, well developed transmission and distribution networks and a favourable fuel price environment. This facilitated electricity reform by allowing for a signiﬁcant period of uncertainty and learning where little net new investment was necessary. (Though in many of the early reformers there was signiﬁcant new investment – Chile and Argentina being the most striking examples in terms of demand growth).

EBL

Foreword: Liberalization and Regulation in Electricity Systems

xxxi

However the investment demands, even on mature networks are now increasing as networks require replacement (or refurbishment) and signiﬁcant new generation investments are required given the investment cycle. This will test the investment incentive structure in these markets (and will test government commitment to leaving investments to the market). In networks, new mechanisms need to be developed for incentivising least cost investments rather than simply incentivising efﬁcient operation (as under RPI-X). Innovative new schemes for the selection of new investments have been successfully tried out – such as the public contest method for new transmission investments in Argentina which involved users voting for whether they wanted new lines (Littlechild and Skerk, 2004). Competitive tendering, even for investments within meshed networks, could be extended with the tender price being incorporated in the regulatory asset base. Correlje and De Vries bring us back to the consistency of electricity reforms with underlying institutional determinants as well as physical and economic factors. This is in line with recent work being done on the determinants of economic growth by La Porta et al. (1999) and the strength of ﬁnancial systems by Bordo (2006). It seems clear that the detailed electricity system reform needs to be consistent with the institutional framework within which the reforms take place. Starting points are important: signiﬁcant public ownership and prices which more than cover efﬁcient economic cost greatly facilitate a structural reform which will yield positive social welfare. These starting points were present in Chile, Argentina, UK, Australia and New Zealand. Initial private ownership (e.g. in the US, Japan and Germany) and prices below economic cost (e.g. in India) make reform much more difﬁcult. Electricity reforms are complex and require a commitment to competition and efﬁcient regulation. Signiﬁcantly they also need to show ﬂexibility to emerging information and allow scope for mid-course adjustments. It seems clear that a jurisdiction like the UK is very well suited to delivering a successful electricity reform at the generation, network and retail levels. This is because the UK has a strong central state capable of encouraging private sector compliance under threat of legislation, a deep commitment to competition and liberalised ﬁnal prices, a tradition of independent regulation and a signiﬁcant capacity for institutional learning. Part of the problem faced by other jurisdictions is that some of the elements of the required institutional capability are not present and hence severely limit the capacity of the society to deliver a successful electricity reform. Another way of putting this, is to say that what we might call the Standard Reform Design as followed by the UK, Texas etc is a model with an institutional “ﬁt” appropriate to those jurisdictions. Of course this is not to say that for a given jurisdiction an alternative and potentially equally successful model, more in keeping with the institutional environment, does not exist. It is merely to say that the Standard Reform Design to which the EU and the other reforming states are implicitly working will not be appropriate in many, if not most, jurisdictions. It is however worth pointing out that the UK’s (and one suspects, that of most other reforming states) institutional capabilities in the area of recent electricity reforms have been recently acquired and have not always been present. That said, it is true that a striking conclusion from a comprehensive survey of the electricity deregulation process in the UK is how impressive the British civil service was in designing and implementing a comprehensive reform, from scratch, in less than two years (Henney, 1994). To come full circle: I think there is an appropriate balancing of liberalisation and appropriate regulation in a given electricity system. However I think that the leading jurisdictions in electricity reform challenge everyone else to justify why reform ought not to be extended into areas previously thought to be unsuited to competition. This challenge seems particularly appropriate in the US where the contrast (at the state level) between leaders and laggards in reform seems marked and without much institutional logic (in the La Porta

EBL

xxxii

Competitive Electricity Markets

et al. and Bordo sense), though one can undoubtedly come up with theories to explain them (the Joskow, 1997, discussion of reform attitude and pre-reform price being the most economically satisfying). In other countries however the argument for the balance to be set in different places becomes easier to justify on the basis of differences in institutional capability. The challenge posed by this is then to improve institutional capability in order to exploit the possible gains from electricity liberalisation. The real prize for proceeding with this is an electricity system capable of fully harnessing the power of competitive forces to respond to an uncertain future for the world’s energy markets in the face of environmental, technological and geopolitical challenges. Acknowledgements This paper is based on a talk to the 3rd Annual Regulation Seminar, SBGI, 15 March 2007. The author is grateful for the ongoing ﬁnancial support of the ESRC Electricity Policy Research Group. The comments of Perry Sioshansi, Paul Joskow, Stephen Littlechild and Andy Ford are gratefully acknowledged. All responsibility remains that of the author. References Averch, H. and Johnson, L.L. (1962). Behavior of the ﬁrm under regulatory constraint. Am. Econ. Rev., 52, 1052–69. Barmack, M., Kahn, E. and Tierney, S. (2007). A cost-beneﬁt assessment of wholesale electricity restructuring and competition in New England. J. Reg. Econ., 31(2), 151–184. Baumol, W.J., Ordover, J.A., and Willig, R.D. (1997). Parity pricing and its Critics: A necessary condition for efﬁciency in the provision of bottleneck services to competitors. Yale. J. Regul., 14(1), 145–63. Berg, S.V. and Jeong, J. (1991). An evaluation of incentive regulation for electric utilities. J.Regul. Econ., 3(1), 45–55. Bergman, L., Neven, D.J., Gual, J., et al. (1998). Monitoring European Deregulation: vol.1: Europe’s Network Industries: Conﬂicting Priorities (Telecommunications), London: Centre for Economic Policy Research. Bergman, L., Newbery, D.M.G., Pollitt, M., et al. (1999). Monitoring European deregulation: vol.2: A European market for electricity? London: Centre for Economic Policy Research. Bertram, G. (2006). Restructuring the New Zealand Electricity Sector 1984–2005. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger eds), Oxford: Elsevier, pp.203–34. Bordo, M. (2006). Sudden Stops, Financial Crises and Original Sin in Emerging Countries: Déjà vu? http://michael.bordo.googlepages.com/SuddenStops05–08.doc. Codognet, M-K. Glachant, J-M. Lévêque F.and Plagnet M-A. (2002). Mergers and acquisitions in the European electricity sector cases and patterns, CERNA, Centre d‘économie industrielle, Ecole Nationale Supérieure des Mines de Paris, August, Paris. Cohen, L.R. and Noll, R. (1991). The Technology Pork Barrel. Washington D.C.: Brookings Institute. Demsetz, H. (1968). Why regulate utilities?. J. Law. Econ., 11, 55–65. Domah, P.D. and Pollitt, M.G. (2001). The restructuring and privatisation of the regional electricity companies in England and Wales: A social cost beneﬁt analysis. Fiscal Studies, 22(1), 107–46. Ellerman, A. and Dubroeucq, F. (2004). The Sources of Emission Reductions: Evidence from U.S. SO2, MIT CEEPR Working Paper, 004–001. European Commission (2007). DG Competition Report on Energy Sector Inquiry. Brussels: European Commission. Fri, R.W. (2003). The role of knowledge: Technological innovation in the energy system, The Energy Journal, 24(4), 51–74. Green, R. (1995). The cost of nuclear power compared with alternatives to the magnox programme. Oxf. Econ. Pap., 47, 513–24.

EBL

Foreword: Liberalization and Regulation in Electricity Systems

xxxiii

Green, R., Lorenzoni, A., Perez, Y. and Pollitt, M. (2005). Policy assessment and good practices. In Sustainable Energy Speciﬁc Support Assessment (SESSA): Conference on implementing the internal market of electricity: proposals and timetables, 9 September 2005, Brussels. Green, R. and McDaniel, T. (1998). Competition in electricity supply: Will “1998” Be Worth It? Fiscal Studies, 19(3), 273–93. Fabrizio, K., Rose, N. and Wolfram, C. (2007). Do markets reduce costs? Assessing the impact of regulatory restructuring on U.S. Electric Generation Efﬁciency. Am. Econ. Rev., 97(4), 1250–77. Hattori, T. and M. Tsutsui (2004). Economic impact of regulatory reforms in the electricity supply industry: A panel data analysis for OECD countries. Energy Policy, 32(6), 823–32. Henderson, P.D. (1977). Two british errors: Their probable size and some possible lessons, Oxf. Econ. Pap., 29(2), 159–205. Henney, A. (1994). A Study of the Privatisation of the Electricity Supply Industry in England and Wales, London: EEE Ltd. Jamasb, T., Mota, R., Newbery, D., and Pollitt, M. (2004). Electricity sector reform in developing countries: A survey of empirical evidence on determinants and performance. Cambridge Working Papers in Economics, No.0439. Jamasb, T. and Pollitt, M. (2001). Benchmarking and regulation: International electricity experience. Utilities Policy, 9(3), 107–30. Jamasb, T. and Pollitt, M. (2005). Electricity market reform in the European Union: review of progress toward liberalization and integration, The Energy Journal, 26(Special Issue), 11–41. Jamasb, T. and Pollitt, M. (2007). Incentive Regulation of Electricity Distribution Networks: Lessons of Experience from Britain, Energy Policy, 35(12), 6163–6187. Joskow, P. L. (1997). Restructuring, Competition and Regulatory Reform in the U.S. Electricity Sector. J. Econ. Perspect., 11(3), 119–38. Joskow, P. L. (2005). Incentive Regulation in Theory and Practice: Electricity Distribution and Transmission Networks, EPRG Working Paper 05/11. Joskow, P. L. (2006a). The Future of Nuclear Power in the United States: Economic and Regulatory Challenges, AEI-Brookings Joint Center for Regulatory Studies, Working Paper 06–25. Joskow, P.L. (2006b). Markets for Power in the United States: An Interim Assessment. The Energy Journal, 27(1), 1–36. Kaserman, D.L. and Mayo, J.W. (1991). The measurement of vertical economies and the efﬁcient structure of the electric utility industry. J. Ind. Econ., 39(5), 483–502. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R. (1999). The quality of government. J. Law. Econ. Org., 15, 222–79. Littlechild, S.C. (2000). Privatization, Competition, and Regulation in the British Electricity Industry, With implications for Developing Countries, Energy Sector Management Assistance Program (ESMAP), February, World Bank. Littlechild, S.C. (2002). Competition in retail electricity supply. Journal des Economists et des Etudes Humaines, Vol.12(2/3), June/September, pp.379–402. Available as CMI Electricity Project Working Paper, No.09. Littlechild, S.C. and Skerk, C.J. (2004). Regulation of transmission expansion in Argentina Part I: State ownership, reform and the Fourth Line, CMI Electricity Project Working Paper, No.61. Mota, R. (2004). Restructuring and Privatisation of Electricity Distribution and Supply Business in Brazil: A Social Cost-Beneﬁt Analysis, CMI Electricity Project Working Paper, No.16. Nemoto, J. and Goto, M. (2004). Technological externalities and economies of vertical integration in the electric utility industry. Inter. J. Ind. Org., 22(1), 67–81. Newbery, D. (2006). Electricity liberalization in Britain and the evolution of market design. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger eds), Oxford: Elsevier, pp. 109–44. Newbery, D.M.G. and Pollitt, M.G. (1997). Restructuring and Privatisation of the CEGB – was it worth it? Journal of Industrial Economics, 45(3), 269–304. New Zealand Ministry of Economic Development (2000). Inquiry into the Electricity Industry, June 2000. http://www.electricityinquiry.govt.nz/reports/ﬁnal/ﬁnal-01.html Ofgem (1999). Reviews of Public Electricity Suppliers 1998 to 2000 – Distribution Price Control Review – Final Proposals, December 1999, London: Ofgem.

EBL

xxxiv

Competitive Electricity Markets

Peltzman, S. (1976). Towards a more general theory of regulation. Journal of Law and Economics, 14, 109–47. Pollitt, M.G. (2004). Electricity reform in chile: Lessons for developing countries. J. Net. Ind., 5(3–4), 221–62. Sioshansi, F.P. and Pfaffenberger, W. (eds.) (2006). Electricity Market Reform: An International Perspective. Oxford: Elsevier. Steiner, F. (2001). Regulation, industry structure and performance in the electricity supply industry. OECD Economic Studies, No. 32. Stigler, G. (1971). The theory of economic regulation. Bell. J. Econ., 2, 3–21. Sweeney, J.L. (2006). California electricity restructuring, the crisis and its aftermath. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger eds), Oxford: Elsevier, pp. 319–82. Thomas, S. (2006). The grin of the Cheshire cat. Energy Policy, 34(15), 1974–83. Toba, N. (2007). Welfare Impacts of Electricity Generation Sector Reform in the Philippines, Energy Policy, 35(12), 6145–6162.

EBL

Preface: Competition and Long-Term Dimensions of Electricity Supply WOLFGANG PFAFFENBERGER Jacobs University Bremen

Has market reform in the ESI been a success story? To judge on this question we need well-deﬁned criteria. Such criteria based on social welfare need to have a time horizon. In this introduction we ask what market opening can contribute to the long-term problems of electricity supply. Availability of energy resources, restrictions of carbon emissions, social preferences for the energy mix are some of the important dimensions of the future of electricity. This preface concentrates on the role competition can play regarding these dimensions and the appropriate regulatory mix to reconcile long-term expectations with short-term efﬁciency gains. The availability of energy, especially electricity, is one of the basic preconditions of modern life. Price development in recent years has indicated scarcity of energy resources. The future availability of energy for electricity production is one of the central questions for the development of modern society: how much energy in what form and at what price will be available in the future and what technologies will be used for transforming primary energy resources into electricity? Market opening in the ESI was basically driven by the wish to increase short-term beneﬁts for consumers. This preface asks how market opening and competition can contribute to solving the longer-term problems as described in the ﬁrst paragraph. This exposition is intended to ask a number of questions regarding future developments and the view taken is strongly inﬂuenced by experience with market opening in central Europe. Market restructuring on the one hand and the solutions developed for the longer-term problem of securing energy supply on the other have to be taken together. Competition in the ESI has been promoted by legislators in many countries to

EBL

• •

raise the industry’s operational and allocative efﬁciencies; and promote internal and external trade in the industry to make use of comparative advantages between different market participants, different regions, and countries.

Although there are many differences between market opening approaches in different countries due to their historically given institutional and political structures, a lot of similarities can be found.1 How successful are market reform measures? To ﬁnd out about the success of institutional reforms we need to have an idea of the performance of a properly functioning competitive market and use it as a benchmark for the performance of the industry in its present market setup. This is not a trivial task because it is difﬁcult to interpret market 1 See Joskow, P. (2006). Introduction to electricity sector liberalization: Lessons learned from cross-country studies. In Electricity Market Reform: An International Perspective (F.P. Sioshansi, and W. Pfaffenberger, eds), Amsterdam, pp. 1–32.

xxxv

xxxvi

Competitive Electricity Markets

results unambiguously as a result of the prevailing market rules. In fact, it is almost impossible to give a deﬁnition of competition. Competition is an open process, which produces unexpected results. This understanding of markets goes back to Hayek and his idea of markets as a spontaneous order: The key idea of the spontaneous-order thesis is that self-organizing and self-replicating structures arise without even the possibility of design. To simulate the efﬁciency of the market order by central planning is impossible for epistemic reasons. The norm system underlying the extended order of a free society was not designed, not invented, and only retrospectively can we recognize it as the condition for retaining at least what we have got.2 Thus to design a market is a contradiction in itself. Typically, markets are created by actors within the market, and not creators or designers from outside. “The market is not simply ‘there’, given by nature, but has to be formed or produced. The producers of a market (‘market makers’) are themselves market actors.” 3 Is this general principle also applicable in electricity markets, where outsiders deﬁne and design the market and establish the rules and then the players function according to the rules laid out? One of the general principles of competition policy therefore has always been to deﬁne conditions that do not allow the development of competition rather than trying to give a positive deﬁnition of what competition is or should be. The expectations regarding the effects of competition on the short-term performance of the industry are often high and political intervention is often based on such expectations. Interventions into market rules are caused by such short-term considerations and often promised results for consumer beneﬁts cannot be achieved. The interface between industry and state intervention is a fundamental problem of a market economy in a democratic society. In the words of Furubotn and Richter:

EBL

The decisive difference between economic and political competition is that economic competition is the struggle for economic advantages through economic exchange. It occurs on the basis of secure property rights. Political competition, on the other hand, is the struggle for authority, that is, the power to change exactly these property rights – unilaterally without any economic quid pro quo. Thus, in politics, property rights are “up for grabs,” not only economic rights but also …“political property rights,” the rights to exercise public authority. As a result, institutional frameworks may change frequently in a democracy and thus may be quite unstable.4 If you take the view of economic actors in the playing ﬁeld, you will come to different conclusions regarding the desired institutional setup than if you take the view of a government feeling pressed to present successful interventions to placate to the short-term expectations of voters, most of whom are small-scale consumers. Moreover, the time horizon of the typical legislator is often much shorter than the lifetime of physical assets in the energy industry. Economists are brought up in the spirit of free markets and are generally skeptical of government intervention. Their advice usually is to leave markets alone, assuming that 2

Radnitzky, G. (1984). Die ungeplante Gesellschaft – Friedrich von Hayeks Theorie der Evolution spontaner Ordnungen und selbstorganisierender Systeme. Hamburger Jahrbuch für Wirtschafts- und Gesellschaftspolitik, 29, 9 ff. 3 Furubotn, E. and Richter, R. (2005). Institutions and economic theory, 2nd ed., Ann Arbor, 314. 4 Furubotn, E. and Richter, R. (2005). Institutions and economic theory, 2nd ed., Ann Arbor, 482.

Preface: Competition and Long-Term Dimensions of Electricity Supply

xxxvii

they are competitive, fair, and transparent. Due to the network orientation of the ESI, competition is, of course, not possible in all segments of the industry. In the ESI, transport and distribution constitute a monopolistic bottleneck for generation and supply and therefore need to be regulated. Competition in generation and supply depends on regulated network access at fair conditions for all market participants. Due to the physical integration of production, network, and supply, the regulation in transport and distribution will also have consequences for the potentially competitive parts of the system in generation and supply. Therefore regulators may be tempted to also look at the competitive parts of the industry. This, in addition to the different historically grown institutions in many countries, is the reason for different approaches to competition in this industry. Basically, we can distinguish two concepts: The ﬁrst, free market with minimum regulation, puts the emphasis on the freedom of the market and tries to restrict regulation to the minimum required to allow competition5 ; the second, market results for price and investment, puts the emphasis on regulation and basically wants to achieve certain market results with regard to price development and security of supply issues particularly regarding investment in generation. The two concepts are theoretical extremes and in the real world we can ﬁnd a mixture of elements of both concepts.

Free Market with Minimum Regulation

EBL

A free market requires privatization of public companies, an appropriate form of vertical separation of transmission and distribution and ancillary services from the competitive segments and free network access at fair conditions. In addition, all the usual control instruments against the misuse of market power need to exist and to be used. It is debatable whether network prices need to be regulated explicitly or implicitly by using the instruments against the misuse of market power on a case-by-case basis. In the second case, the effects may be as good as in the case of full-scale regulation but at much lower bureaucratic cost. Vertical integration in the power industry is the result of economic incentives because specialized assets in generation, transmission, and distribution need to be combined physically and vertical integration is the answer to reduce risk and transaction cost.6 Full vertical integration, however, is an obstacle to competition so that some form of disintegration is necessary for competition. The tradeoff between vertical integration and unbundling does not automatically favor a full-scale unbundling,7 including ownership unbundling as presently envisioned by the European Commission for the European power market. In a free market we expect that the market, to a large extent, invents itself and also develops rules for interaction as part of the game and through contractual relationships. This can only take effect, however, if government and legislation restrict their activity to the absolute minimum as described above. 5 See Knieps, G. (2006). Sector speciﬁc market power regulation versus general competition law: Criteria for judging competitive versus regulated markets. In Electricity Market Reform: An International Perspective (F.P. Sioshansi, and W. Pfaffenberger, eds), Amsterdam, pp. 49–74. 6 Douma, S. and Schreuder, H. (2003). Economic approaches to organizations, 3rd ed., Harlow, 203. 7 See Chao, H., Oren, S., and Wilson, R. (this book). Reevaluation of Vertical Integration und Unbundling in Restructured Electricity Markets. Chapter 1.

xxxviii

Competitive Electricity Markets

Market Price and Investment In the time before liberalization of the energy markets when vertically integrated companies prevailed, energy services were often considered an infrastructure good so that a high level of public interest and regulation was the countervailing power to the market power of the local or regional vertically integrated monopolies. After market opening, the focus of regulation should shift to supporting the development of competition by adequate institutions and not look at day to day market results. This is often not the case. Regulation often goes far beyond the required minimum for the market to function. In a free market one would expect that generators would choose the mix of technology and primary energy they ﬁnd suitable to minimize price and quantity risks. In addition, one would expect them to decide on the capacity they ﬁnd adequate. On the other hand, the public has views on the energy mix and may restrict the choice of technology/energy and may want a higher security of supply so that a certain control of investment may be established. Governments may also want to protect consumers against price risks. Such a protective regulation hardly ﬁts into a market environment and is a contradiction in itself. Such a regulation which had its reason in the time of development of the industry and was meant to allow universal access to the services of the industry during its infancy is outdated when the diffusion of the product has reached a high level. We observe however, that political institutions often live much longer than their intended purpose. It is a basic principle of a market economy that those who bear the risks should have the freedom to make decisions. This is true for producers and consumers assuming that the marketplace allows enough possibilities of choice. There is a general tension between intervention and market and this is particularly important in the energy industry due to its special features. Regulatory systems in different countries often represent a mix of the basic concepts of competition and the pure market model can hardly be found. In the following a few key questions are posed regarding the implementation of the market model.

EBL

Choice of Primary Energy Can markets with a high degree of regulation reach an energy mix favorable for customers? Sources for low-cost energies are restricted partly by natural factors and partly by political decision, e.g., in the case of nuclear energy. The implication of this can be seen from Fig. 1. Whereas the United Kingdom is often considered as a functional model for a competitive market (e.g., low degree of market concentration in generation, ownership unbundling in transmission), its performance in recent years as measured by the wholesale prices is much worse than in central European countries. This is partly due to the insular situation and mainly a result of the power mix with a high share of natural gas and an attempt to escape the rising gas prices by going back to coal, which resulted in a net import of carbon certiﬁcates at high prices at certain times. A perfect market model and short-termoriented optimization does not automatically lead to better performance. So what kind of freedom do utility companies need to reduce energy price and quantity risks in the longer term? Do they need a more long-term orientation and in what relation is this to the time horizon of policymakers? There deﬁnitely is a tradeoff between freedom in the market and regulation.8 8

This topic is also addressed by Michael Pollitt, Foreword, this volume.

Preface: Competition and Long-Term Dimensions of Electricity Supply

xxxix

Wholesale prices front year base 90

France

UK 80

Netherlands Germany

EUR/MWh

70

60

Nordic

Belgium

50

40

30

Source: EST 20 10.1.03

10.5.03

10.9.03

10.1.04

10.5.04

10.9.04

10.1.05

10.5.05

10.9.05

10.1.06

10.5.06

10.9.06

10.1.07

EBL

Fig. 1. European wholesale prices, Source: EST (2007).

Market Integration In Fig. 1, the Nordic countries, which are only partly connected to the central European grid, have a considerably lower price level than the rest of Europe. This can be explained to a large extent by the energy mix with a high share of low-cost hydro and nuclear. A higher degree of integration with the rest of Europe through higher transmission capacities would lead to price adjustment between the Nordic and the central European market. This could be in the interest of the producing companies in the Nordic market as they could earn additional proﬁts because marginal capacities in central Europe would set the price in the market. But is this in the interest of Nordic consumers or governments? In the longer run, according to the standard economic paradigm, the comparative advantage of the low-cost producers should lead to an increase in low-cost capacities. Prices would adjust to the level of low-cost producers and capacities would migrate from high-cost producers to low-cost producers. But is this likely to happen? In reality this is highly unrealistic because low-cost production possibilities are scarce due to many, often political, restrictions. Taking this into account what will be the optimum market integration and to what extent should transmission capacities be expanded? Generally, the open market approach only makes sense if enough freedom of choice exists so that considerable reallocation effects can be expected. If there are too many restrictions in the system, market opening will lead to redistribution without reallocation. To give an example: Hydro power producers in Austria are happy to supply German networks with reserve power to compensate for ﬂuctuations of wind energy at an increasing rate. This leads to windfall proﬁts in Austria because the marginal capacity in Germany is coal-ﬁred and has a much higher variable cost. From the point of view of allocation this makes

xl

Competitive Electricity Markets

sense, but will this increase the acceptance of markets in public and in the political system if consumer prices in Austria are also rising? Renewable Energy The promotion of renewable energy as a substitute for fossil fuels in many countries is often achieved outside of the market mechanism, because the willingness of consumers to pay for the quality differential is not sufﬁcient to compensate for the higher cost of many renewable options, at least at current prices. After a certain level of renewable penetration has been reached, this of course distorts the market process if one section of the market lives behind a protective fence and the other has to adjust to whatever happens behind that fence. This is particularly true if renewable production is stochastic as in the case of wind. Conventional producers have to adjust their output and portfolio to the dynamics of the renewable generators without adequate compensation, whereas the production of renewables is often kept free of market risks. The question then arises: what is the best approach for market integration of renewables regarding ﬁnancial support and system integration?9 There are two reasons for the promotion of renewables: Substitution of scarce fossil resources and reduction of greenhouse gas emissions. The question is, whether the signals given to the market by the political instruments for promoting renewables are consistent with these goals and the investment incentives or disincentives for conventional generation are consistent with the necessary backup function of the conventional generation section. This can only be achieved if reserve power needed for stabilizing renewable generation is integrated into the scheme of promotion. In Europe the now existing system of CO2 certiﬁcates in itself gives signals to the market regarding the portfolio of desired generation based on the carbon content of fuels used. This leads to a relative advantage for carbon-free renewables because the production cost of fuels with carbon content will rise depending on the quantity of certiﬁcates allocated to electricity generation. Again, the question to ask is whether this leads to a set of signals consistent with the other signals produced by promotion schemes for renewables. Assuming that the introduction of competition in the ESI leads to efﬁciency gains, it is quite likely that the economic and ﬁnancial implications of support schemes for renewables are reaching levels that become much higher than any efﬁciency gains that can be reached through competition in the core of the electricity market. The economic relevance of competition is reduced unless schemes of promotion are made to ﬁt much better into the competitive market. Increasing the share of renewables to a relevant size like the envisaged share of 20% in the European Union (EU) in the future will lead to a number of new questions regarding the market environment as a whole.

EBL

Miracle Competition? Intensity of competition depends strongly on the maturity of a market. In a mature market with low growth and high economies of scale, as is typical for electricity generation in developed countries, competition between existing generators will make it difﬁcult for newcomers to enter the market. Existing generators with a historically grown portfolio of new (expensive) and old (cheap) plants and a developed set of customer relations 9

See Chapter by Haas et al., Chapter 12, this volume, for further details.

Preface: Competition and Long-Term Dimensions of Electricity Supply

xli

can reduce the risks considerably. Thus the threshold for intensive competition may be high. But does it make sense to increase the number of generators by separating existing companies and thereby increasing the risks and transaction cost?

Customer Switching Electricity is a commodity, and one with few opportunities for differentiation or added value. Still, customers may develop preferences for certain suppliers due to conﬁdence or added services. Customer behavior, particularly by small customers, is strongly inﬂuenced by the metering and billing services and the ease of available payment options. For most consumer goods, market extraction and payment are synchronized, so that price and expenditure are experienced in the context of market extraction. In the electricity market there is no such direct connection between consumption and payment. Price and consumption are therefore not directly related as is true for most other consumer goods. It is often said that the intensity of competition can be measured by the number of customers who switch suppliers. This is certainly true for industrial and commercial companies with a professional purchasing department. It may, however, not be true for residential customers. Consumers tend to develop habits and use certain suppliers regularly. A signiﬁcant price differential may be needed to change that habit – which is difﬁcult to imagine for smaller consumers. If the competitive advantage is smaller than the margin required to make the consumer move, little switching will occur. From the point of view of the supplier, it makes sense to try to ﬁnd out about that margin and keep the supply price within that margin against competitive offers. The stability arising from such a behavior certainly is a result of competition. Low switching rates are not an unambiguous proof of low degree of competition. It rather depends on availability of access and transparency of the market. After all, certain price differences can be found in all markets for goods with homogenous quality. Compare the prices for branded goods between two adjacent supermarkets and you will ﬁnd a lot of differences; still, all items in both markets will sell. This can be explained by imperfect information, transaction costs as well as bounded rationality.10 We should not expect the “perfect consumer” in the power market. In the power market for residential customers it is only the competitive part of total price that matters for competition. Depending on the level of consumption, the regulated network fees, taxes, and levies, this competitive fraction of the price may be relatively small and below the margin that would induce customer switching. The German example may be extreme. For the average residential customer, after deducting taxes, levies, and regulated network fees, the competitive part of the power price is about 25% of the total. In absolute terms, a residential customer can earn $5–10 per year by switching suppliers, probably below the threshold for switching for most consumers. It is, however, necessary to have open competition so that suppliers are under pressure from potential competition; on the other hand the competitive advantage will always be relatively small for smallish consumers. How relevant then is competition in the small retail market for social welfare? Is it not the wholesale market that really matters because the advantage of reallocation and trade within the system as a whole leads to a relevant size of efﬁciency gains?11

EBL

10

See Douma, S. and Schreuder, H. (2003). Economic Approaches to Organizations, 3rd ed., Harlow, chapter 4 and chapter 8. 11 See also Joskow, P. (2006). Electricity Market Reform, p. 23.

xlii

Competitive Electricity Markets

Unbundling Nearly everyone agrees that a certain degree of vertical separation is necessary to allow competition. But how far should it go and should it be physical or functional? The European Commission is proposing a complete ownership unbundling for EU member countries. The unmentioned but real agenda behind this policy seems to be a desire to develop a pan European network that would be much easier to unify if national gridcos are independent from all other interests and can be merged much more easily into a pan European company. The Commission seems to think that such a scheme would help in addressing some pressing matters such as the future availability of primary energy import, help to reduce energy cost, and support European economic development. Cooperation between European grids has worked quite well in the past, there deﬁnitely will be new challenges for the grids in the future with rising energy trade, the increasing share of renewables and rising distance between point of production and consumption. This is a long-run problem that has lots of highly technical implications. Gridcos and regulators have to cooperate to develop the necessary solutions. But it seems difﬁcult to see these future challenges being met without a clear bottom-up approach. Playing around with institutions from a central perspective might do more harm and delay the necessary development. Conclusion Around the year 2000 the paradise of abundant and relatively cheap energies came to an end. At the same time the industrial world has to accept the challenges of becoming more import dependent and having to restrict the use of fossil fuels to achieve greenhouse gas reductions. This implies a lot of fundamental changes for primary energies and technologies to be used, for the role of distributed generation, for network operation and development. Competition is an important element to achieve the changes required. But there seems to be a tradeoff between the short-term goals of customer beneﬁt and the long-term goal of promotion of structural change including smaller share of fossil fuels, less greenhouse gas emissions. For the latter, investment incentive is the main topic. New products and technologies require adequate prices. Short-term gains maybe long-term losses! In the long term, governments have to deﬁne clear goals regarding the desired energy mix and environmental targets. Markets can then work to achieve these goals if they are left with the necessary freedom to experiment with technology and earn rates of return including the risks for innovative action. The power market is not an isolated market. We have to consider it as part of the whole industry. By regulating the power market too tightly, we also indirectly regulate all the markets that produce speciﬁc assets for the power industry. This may not be compatible with long-term goals.

EBL

Introduction: Electricity Market Reform – Progress and Remaining Challenges FEREIDOON P. SIOSHANSI Menlo Energy Economics

Summary For over two decades, policymakers and regulators in a number of countries around the world have been grappling with market reform, liberalization, restructuring, and privatization of the electric power sector. While a great deal has been achieved – including many useful lessons in what works, what does not, and why – successful design and implementation of market reform still remains partly art. Moreover, the international experience to date indicates that, in nearly all cases, initial market reform leads to unintended and occasionally unpleasant consequences or introduces new concerns and risks that must be addressed in subsequent “reform of the reforms.” This introduction provides an overview of the market reform experience, identiﬁes a number of challenging issues facing regulators and policymakers, and provides the context for the remaining chapters of the book, which cover these topics in detail.

EBL

Electricity Market Reform: An Overview Since the mid-1980s, a number of countries around the world have engaged in market reform initiatives including liberalization, privatization, and/or restructuring the electricity supply industry (ESI). Box 1 describes terminology of different approaches to market reform. The motivations for changing the organization of the industry and the regulatory paradigm vary from case to case, but are generally driven by a desire to introduce competition in the hope of making the industry more efﬁcient, making prices more transparent, and transferring more risks to private investors rather than ratepayers or taxpayers. The ensuing productivity improvements, better rationalization of labor and fuel costs, superior choice of generation technologies, allocation of investment risks, and other measures are believed to lead to lower electricity costs and improved services beneﬁting ultimate consumers. Chile is generally credited as the ﬁrst country to undertake a major market reform process in 1987 (Raineri, 2006). England and Wales followed with a privatization and liberalization scheme in 1989 – which has been widely studied and copied (Newbery, 2006). Since then, the pattern has been repeated in a number of countries large and small (Table 1). 1

2

Competitive Electricity Markets

Box 1 Taxonomy of key market reform terms Terminology used to describe different approaches to change the regulatory paradigm Restructuring is a broad term, referring to attempts to reorganize the roles of the market players and the regulator, and/or redeﬁne the rules of the game, but not necessarily “deregulate” the market. California, for example, restructured its market, “deregulated” its wholesale market by lifting nearly all restrictions to how wholesale prices could be set by generators, but kept its retail market fully “regulated,” in this case capped. Many problems ensued. Liberalization is synonymous with restructuring. It refers to attempts to introduce competition in some or all segments of the market, and remove barriers to trade and exchange. The European Union (EU), for example, refers to their efforts under this umbrella term. Privatization generally refers to selling government-owned assets to the private sector, as was done in Victoria, Australia, with former SECV, in Italy with ENEL, and in France with EdF and GdF. It must be noted that one can liberalize the market without necessarily privatizing the industry, as has successfully been done in Norway and in New South Wales, in Australia. Corporatization generally refers to attempts to make state-owned enterprises (SOEs) to look, act, and behave as if they were for-proﬁt, private entities. In this case, an SOE is made into a corporation with the government treasury as the single shareholder. For example, former SOEs in New South Wales (NSW), Australia, have been corporatized. They vigorously compete with one another, while all belong to the same, single shareholder, namely the Government of NSW. Deregulation is essentially a misnomer. No electricity market has been (or, in fact, can be) fully deregulated. Experience suggests that even well-functioning competitive markets need a regulator, or as a minimum, a market monitoring and anticartel authority.∗ Until recently, Germany was the only major country attempting to go without a regulator but with an anticartel ofﬁce, monitoring the behavior of the market participants.

EBL

Source: Sioshansi, F. and Pfaffenberger, W. (2006). Electricity Market Reform: An International Perspective. Elsevier. ∗ The issue of how much regulation is needed and in what form is an open-ended issue. Most economists agree that electricity markets, at the minimum, need a market operator or manger to regulate the ﬂow of power. There is also a general consensus that electricity markets need a market monitor to prevent monopolistic or opportunistic behavior. Most electricity markets also have a regulator to enforce the laws and the rules and to continue regulating the network components of business, which remain regulated even after liberalization. There is little agreement on what is the proper role and style of regulator because of ideological and political differences of opinion. Some economists do not see any need for a powerful and/or interventionist regulator any more than such a regulator is needed for the auto or tire industry.

Introduction

3

Table 1. Selected countries with electricity market reform Country

Market reform highlights and comments

Argentina

Has experienced problems due to external economic crises.

Australia

Various dates in various states, some states have privatized, others corporatized, resulting in uneven playing ﬁeld. Generally considered a successful market despite some remaining problems further described in this volume.

Brazil

Initial market reforms continue to be modiﬁed due to problems with initial market design, further covered in this volume.

Canada

Alberta and Ontario introduced competition. Ontario has rescinded it following public pressure after price increases; Alberta has succeeded, progress stalled in other provinces.

Chile

First to introduce market reform in 1987, continues to evolve.

Colombia

Introduced market reform in 1994–95, is experiencing problems in retail, wholesale, and capacity markets, reform of reforms in progress.

England and Wales

Introduced radical privatization and restructuring in 1989, has gone through at least three major phases of reform, continues to evolve, has been widely studied and copied as a successful model.

European Union

25 members of EU continue to make slow progress, various deadlines for unbundling and introduction of retail competition have been set but not fully implemented, full retail competition was introduced in July 2007 but has not signiﬁcantly changed the status quo, the goal of a fully integrated pan-European market remains elusive despite the best efforts of the EC policymakers.

EBL

Japan

Has introduced limited competition to date with a cautious pace, Japan Electric Power Exchange (JEPX) is in place but there is limited volume of trading.

South Korea*

Has created Korea Power Exchange (KPX) and broken up KEPCO into several generation companies. Full liberalization, however, has stalled due to political and labor opposition.

New Zealand

Experienced some problems in the absence of a regulator, which has been introduced; has a rather complicated nodal pricing scheme.

Nordic countries

Considered among the most successful markets based on bilateral trading, has expanded to include all Nordic countries, market survived a major drought without meltdown.

Singapore*

Considered successful despite a small market, limited number of players, and a complicated nodal pricing regime.

Thailand*

Liberalization and privatization stalled due to labor union opposition and lack of political support.

United States

Wholesale competition encouraged since 1992 with passage of Energy Policy Act and considered successful; retail competition introduced in selected markets since 1998 with mixed results, progress stalled after California electricity crisis in 2000–01, no retail progress since opening of Texas market in 2002.

Note: * Not covered in Electricity Market Reform: An International Perspective (2006: Elsevier). Source: Compiled from various chapters of Sioshansi, F. and Pfaffenberger, W. (2006). Electricity Market Reform: An International Perspective, Elsevier; and other sources.

4

Competitive Electricity Markets

Even when the initial design of the reformed market is sound, the implementation and transition process can go astray, sometimes with serious consequences. This may be among the reasons why market reform initiatives have all but stalled in the United States following the much-publicized failure of the California market in 2000–01. Faced with these uncertainties and risks, the introduction of market reform in South Korea and Thailand, to name a few examples, has stalled. The politicians in both countries face signiﬁcant opposition from the unions and others who want to maintain the status quo. In the meantime, market reform proponents cannot guarantee tangible and immediate beneﬁts that make it politically untenable to push for market reform. In some cases, such as in Thailand or India, the status quo, while not efﬁcient, offers signiﬁcant price subsidies to key segments of the population. The recipients of these subsidies do not favor a liberalized market that would most likely eliminate the subsidies at the ﬁrst opportunity. In the case of Thailand, one study reportedly concluded that to attract sufﬁcient foreign investment, the country’s average wholesale prices would have to be raised by 25% or more. Such a price increase, while necessary to attract foreign investment, would hardly be popular with the voters. Likewise, the liberalization of the power sector in India, another country with massive need for new infrastructure investment, is beset by chronic price subsidies to large segments of population and hampered by institutional obstacles. In the case of Japan, which has one of the highest electricity prices among developing countries, policymakers have decided to proceed incrementally and at a cautious pace (Goto and Yajima, 2006). The risk of unsettling the status quo, which in the case of Japan provides high levels of reliability at a stable but high price, is among the reasons cited for the government’s measured approach. The book Electricity Market Reform: An International Perspective (Sioshansi and Pfaffenberger, eds, 2006) provides an overview of market reform initiatives in a number of countries around the world and draws useful comparisons of what appears to be working, what does not, and why. Among other things, the book points out some of the major lessons that have been learned, sometimes at great expense, from various schemes to introduce market reform in the ESI. While much has been learned in the process and a blueprint has emerged (Joskow, 2006a), market reform still remains partly art and not all science. Moreover, the international experience to date indicates that in most cases, initial market reform leads to unintended consequences that must be addressed in subsequent “reform of the reforms,” to borrow from (Joskow, 2006a). Aside from this, market reform in many countries has introduced new problems, risks, and concerns, some of which are not fully resolved. Addressing some of these remaining issues is the main objective of the present volume.

EBL

Market Design, Implementation Challenges, and Reform of the Reforms The introduction of market reform generally – but not necessarily – follows something along the following lines: The process usually starts with an acknowledgment of problems and/or deﬁciencies associated with the existing system – prompting a search for alternative ways to organize, operate, manage, and regulate the ESI. Motivation to ﬁx real or perceived problems is usually prompted by gross inefﬁciencies in operation and/or performance, poor system reliability, high prices, supply inadequacies, inadequate investment in infrastructure, or a combination of these.

Introduction

5

The old adage, “If it ain’t broke, don’t ﬁx it,” generally applies in the sense that if retail prices are low, supplies are adequate, and the system operates reliably and with tolerable efﬁciency, there is probably no compelling motivation to change. This is evident in the United States, where market reform initiatives have been largely conﬁned to states with above-average costs. Lower-cost states have, by and large, stayed on the sidelines so far and do not appear inclined to introduce market reform initiatives. The second stage usually starts with a debate on what is broke, and how best to ﬁx it, i.e., the process leading to an alternative, hopefully superior, market design resulting in improved performance. With more experience gained over the past two decades, there is now a growing body of knowledge on the main features of market reform (e.g., Joskow, 2006a). In most cases, new laws must be passed and the organization of the ESI must be changed before a new market structure can be implemented. Typically, new institutions, such as a market operator and/or a power exchange as well as a new regulatory framework, must be created, usually at signiﬁcant costs. There are risks in changing the status quo, and the skeptics or those who stand to lose from the market reform process have to be persuaded to go along. Winners and losers must reconcile their gains and pains. In developing countries, new institutional reforms are generally needed before market restructuring can be introduced. The third stage involves the actual implementation of the market design devised in the previous step. While the theory may be sound, there is no guarantee that the implementation will go smoothly or without major hiccups. In many cases, technically sound principles do not translate well in practice resulting in problems and chaos. For example, the introduction of retail competition to large numbers of residential and small commercial consumers has proven problematic logistically in nearly all markets. Likewise, the organization and operation of centrally dispatched market operators and various markets are not trivial. The fourth stage begins with the realization that the introduction of market reform does not necessarily or automatically lead to many of the expected beneﬁts and outcomes. For example, while wholesale auctions generally result in vigorous competition among generators, this need not necessarily result in immediate lower retail costs for consumers. Even when average wholesale prices decline, prices typically become more volatile, creating demand for risk hedging strategies. Nor is there a guarantee that competition will attract additional investment to the electric power sector when there are signiﬁcant regulatory or political uncertainties, or when politicians routinely interfere in key decisions as pointed out by Pfaffenberger in the Preface to this volume. Resource adequacy (RA) is identiﬁed as a concern in many reformed markets. In most cases, the initial market design may have inherent ﬂaws that only become apparent after the passage of some time or during the course of an external event – be it a serious drought, a heat wave, currency devaluation, economic slump, or otherwise. These events occasionally derail what could otherwise have been an acceptable outcome. Fuel price increases unrelated to the market reform process, for example, can result in signiﬁcantly higher retail prices, eroding the political support for market reform.1 In nearly all cases, initial market reform has led to unforeseen and unintended consequences that must be addressed in subsequent “reform of the reforms.”

EBL

1 In the United States, restructuring in a number of states is under attack because the initial reform process legislated a rollback in retail prices, which were frozen for a period of time. The expiration of these frozen rates has resulted in public resentment for the restructuring process.

6

Competitive Electricity Markets

The ﬁfth and ﬁnal stage, usually an ongoing process, involves dealing with the problems associated with the initial market design ﬂaws, ﬂaws in implementation, or unanticipated problems resulting from external factors or events. Droughts, fuel price increases, market dominance of few key players, uneven competition resulting from an uneven playing ﬁeld, poor and/or ineffective market monitoring and regulation, and inadequate investment in infrastructure are among the reasons often cited for the need for subsequent political or regulatory meddling in the market. Partly for these reasons, the introduction of market reform has been eventful with decidedly mixed results internationally. In a few cases, notably in California, well-publicized disasters have dampened interest in neighboring regions or states. In some cases, politicians were able to weather initial kinks, allowing markets to do what they are supposed to do, as in Alberta, Canada. In other cases, as in the Province of Ontario, the politicians withdrew their support of market reform at the ﬁrst sign of trouble – allowing a temporary aberration to derail the process before it had a chance to succeed (Treblicock and Hrab, 2006). By contrast, the Nordic market was able to withstand a signiﬁcant drought with serious hardships but no market manipulation, no blackouts, and no collapse, reinforcing everyone’s faith in the market (Amundsen et al., 2006). Overall, the implementation of market reform has not been easy. The British market, for example, has arguably gone through at least three distinct reform stages, from the original, central, mandatory Pool in 1989 to the New Electricity Trading Arrangements (NETA) introduced in 2001, to the British Electricity Trading and Transmission Arrangements (BETTA), introduced in 2003. Newbery (2006) provides a summary of the evolution of the British market since inception. Likewise, Brazil has gone through several reincarnations of its original reform agenda, each addressing problems introduced by the previous reforms (Araujo, 2006). De Araújo et al. (Chapter 15) use Brazil as a case study of “the reform of the reforms,” describing the evolution of the market over time, including persistent problems that are still being addressed. The process of modifying and reﬁning the original reforms continues in most countries and states including Chile (Raineri, 2006), Argentina, and Colombia (Dyner et al., 2006), and California (Sweeney, 2006), just to mention a few. Broadly speaking, the implementation of market reform has been more manageable in smallish, isolated markets. Large interconnected markets such as those in the United States and continental Europe have proven more challenging. In the latter case, separation of generation from the grid and supply has not been satisfactorily achieved, creating opportunities for some players to subsidize competitive functions from network-based regulated functions (Haas et al., 2006). Likewise, in some markets, notably Australia, some players remain state-owned while others have been privatized, resulting in an uneven playing ﬁeld.2 In the European context, despite numerous European Commission (EC) directives with speciﬁc deadlines, market liberalization has stalled in a number of key countries, and there is little meaningful competition at the retail level in some countries. While a number of European countries have liberalized both electricity and gas markets, others have done so

EBL

2 Some experts do not regard the continued existence of public/private ownership as too serious of a problem. Moran, for example, believes that despite the distortion to the level playing ﬁeld introduced by mixed ownership, “the market and its investment pattern has worked well to date” – at least in the Australian context. But even he admits that, “There are clear vulnerabilities where public ownership decisions are taken on the basis of matters other than commerciality.” (Refer to Chapter 11 by Moran and Skinner)

Introduction

7

only on paper, and even that has been accomplished grudgingly. The introduction of full market liberalization in the EU, which went into effect in July 2007, is viewed as a mere formality by some skeptics. In a survey article on European energy markets, The Economist (2006) identiﬁed the following three as the prime reasons for failure to create a competitive pan-European energy market: •

Deliberate state interference motivated by a desire to support so-called national energy champions. • Lack of interest by dominant players or governments to build additional transmission lines to facilitate cross-border trade. • Weak enforcement of EU directives at the country level. Needless to say, the EU’s long-term goal of creating a harmonious, fully integrated panEuropean energy market has not yet materialized. Instead, Europe has evolved into several distinct submarkets separated by chronic transmission bottlenecks, as in Fig. 1 (Haas et al., 2006). The result is continued retail price disparities and dominant players maintaining signiﬁcant market power and market share in key markets (Fig. 2). Cornwall describes current EU efforts to create a fully integrated network in Europe3 . Similar obstacles frustrate regulators at the Federal Energy Regulatory Commission (FERC) in the United States where progress is being made, albeit at a slow pace.4 In this case, FERC continues to push for the creation of a number of large independent system operators (ISOs), a critical step in introducing competition at the wholesale level and a precursor to competition at the retail level (O’Neill et al., 2006). More than half of the US generation capacity is now under control of a handful of large ISOs – with signiﬁcant implications for retail markets (Table 2, and Fig. 3).

EBL

Europe 2

1

5

3 4

Fig. 1. Electricity sub-markets in Western Europe. Source: R. Haas, Vienna Technology University.

3 4

See chapter in this volume. Chapter 4 by Singh and Chapter 5 by Helman et al. provide further details.

Competitive Electricity Markets

8

Fig. 2. Electricity and gas prices in selected European countries, E ¢/kWh*. Source: The Economist, 11 February 2006 based on NUS Consulting results. ∗ Electricity prices based on April 2005 data for consumer with 1000 kW demand using 450,000 kWh/mo, excluding VAT. Natural gas prices based on September 2005 data for consumer using equivalent of 2,931,000 kWh/yr excluding VAT. Including VAT, which varies from country to country, would signiﬁcantly add to costs.

Table 2. US wholesale markets under ISO or RTO control, 2005 System operator

EBL

Generating capacity (MW)

ISO New England (RTO) New York ISO PJM (expanded) (RTO) Midwest ISO (MISO) California ISO (CAISO) ERCOT (Texas) Southwest Power Pool (RTO)* ISO/RTO Total

31,000 37,000 164,000 130,000 52,000 78,000 60,000 552,000

Total US generating capacity

970,000

*Organized markets being developed. Source: Joskow, P. (2006b). Markets for power in the United States: An interim assessment. The Energ. J., 27(1).

In a survey article on the state of the US electricity market, Joskow concludes that “signiﬁcant progress has been made on the wholesale competition front,” but adds, “The framework for retail competition has been less successful” (Joskow, 2006b). A study of the wholesale and retail markets by FERC came to similar conclusions (FERC, 2006). In the United States, the development of well-functioning, competitive power markets continues to be a work in progress.5 Even though California has soured many state 5 For a comprehensive discussion refer to FERC 2006, Report to Congress on competition in the wholesale and retail markets for electric energy, Washington, D.C., 5 June.

Introduction

9

Fig. 3. Operating ISOs and RTOs in the United States. Source: Report to Congress on competition in the wholesale and retail markets for electric energy, FERC, Washington, D.C., 5 June 2006.

regulators’ appetite for market reform, 20 states and the District of Columbia offer some form of customer choice to all or some of their customers6 (Fig. 4). But there is little doubt that the pace of transition to a national competitive electricity market has stalled in the United States (Fig. 5) for a number of reasons including:

EBL

•

The failure of the California market has left lingering concerns among some statelevel regulators and legislatures. • Mixed results in a number of other states that have introduced retail competition and are now facing “rate shocks” as the original retail price freezes are lifted.7 • Lingering problems in some wholesale markets that have not performed as expected. • Lack of interest by the US Congress to push retail competition at the national level.8 Form, function, and intent of regulation While there is consensus on the need for an independent and competent regulator in competitive electricity markets,9 there is no agreement on how extensive or intrusive this 6

For a discussion of retail markets in the United States see Tschamler, T. (2006). In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. 7 When retail competition was introduced in the mid-1990s in some states, rates were typically rolled back and frozen, in some cases for as long as 10 years. These rate freezes are about to become unfrozen, resulting in rather signiﬁcant price increases in some cases. This has resulted in public discontent in a few states, notably Maryland, where the regulatory commission was essentially dismissed following a signiﬁcant price increase in 2006–7. 8 This lack of interest is evident in the Energy Policy Act of 2005, where the word “retail competition” does not even appear in the massive bill. 9 Germany and New Zealand initially tried to operate their markets without one, but both have subsequently decided that one is needed.

Competitive Electricity Markets

10

Choice for all customers Choice for large customers only No Choice

Fig. 4. Current status of retail competition in the United States. Source: Tschamler, T. (2006). In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier.

EBL 80,000 70,000

6.3 GW

MW Migrated

60,000

6.2 GW 11.9 GW

50,000 40,000

22.3 GW 30,000 20,000 10,000

22.2 GW

0 2001 & prior

2002

2003

2004

2005 YTD

Fig. 5. US load served by competitive suppliers, 2001–05, GW. Source: Tschamler, T. (2006). In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier.

Introduction

11

function ought to be.10 There are differing views on the raison d’etre of the regulator. Some prefer a minimalist role limited to performing the function of a referee, merely enforcing the rules and ensuring fair play. Littlechild, himself an ex-regulator, believes in “Markets when ever possible, regulation when not” (Littlechild, 2006). Others favor a vigilant market monitor or prefer a proactive and interventionist puppeteer11 (see Box 2). While the latter option may seem safe, it is hard to imagine thriving competition coexisting with a proactive and interventionist regulator. Chapter 7 by Adib et al. is focused on the market monitoring function in competitive markets.

Box 2 Alternative views of competition and regulation in the ESI* Fundamentally, there are two extreme views on the form and purpose of regulation •

Create a level playing ﬁeld and referee the game – in this case, the regulator sets the rules in a way that encourages free and fair competition among the players – and plays the role of an impartial referee. • Write a script and make the puppets dance – in this case, the regulator prepares a prescriptive script for the play, making sure that the players act according to the script.

EBL

The central ingredients of the former are well known namely, • • • • • •

•

Remove barriers to entry in generation. Privatize or corporatize all players so they are on the same footing.† Unbundle vertically integrated enterprises to remove cross-subsidies and selfdealing.‡ Ensure that the transmission network is open and accessible to all under transparent and non-discriminatory prices.§ Create wholesale markets that are open and transparent. Ensure that the grid is managed by an independent operator who maintains reliability, manages transmission congestion, and operates various markets to facilitate trade, liquidity, and risk management. Foster competition in the supply business.

∗

The author is indebted to Professor Pfaffenberger for this perspective. In the Australian context, Victoria has done this while neighboring New South Wales has maintained government control of both generation and distribution, creating an uneven playing ﬁeld especially in the retail sector. ‡ Lack of rigorous physical unbundling is considered as a major obstacle to the working of vibrant competitive electricity markets. § This remains a thorny issue in the European and US context. †

10

Politt discusses these issues in this volume’s Foreword. The contributors to Electricity Market Reform: An International Perspective (2006: Elsevier) exhibit varying views on this question. 11

12

Competitive Electricity Markets

These conditions will most likely lead to a number of ﬁrms active in various segments of the business engaging in short- and long-term transactions. The participants will ﬁnd suitable arrangements to transact, and the regulator’s role is primarily that of a vigilant referee, making sure that the rules of the game are obeyed, infringements are caught, and offenders are punished. Otherwise, market participants are free to roam as long as they obey the rules and play within the ﬁeld. This may sound unimaginative, but it is tolerably safe. Most restructured markets have adopted a variation of these two extremes – with varying levels of autonomy and authority for the regulator. This assumes that the players know best, and the market is simply the sum of its components. Free market advocates and those favoring laissez-faire policies favor a limited role for the regulator, allowing market participants maximum ﬂexibility. For example, generators would not be obliged to bid their capacity in the market, nor would they be required to explain their bidding strategies or prices. The latter case assumes that the regulator or market designer knows better and can produce a prescriptive plot for all the players, most likely requiring many rewrites to get it right. In this case, the roles of the players are carefully orchestrated and their moves are remotely controlled, as a dance of puppets. The regulator plays the role of a chorographer in a dance studio, organizing and directing all the moves. If the players do not dance as directed or desired, the choreographer must step in to correct the plot and the script. It may be a lot of fun for a playful and highly imaginative regulator, a scarce combination – perhaps an oxymoron. Most economists would consider this an impossibility, since there is no assurance that the scriptwriter, no matter how clever, can get all the parts right.** Nevertheless, examples of this line of thinking may be found in some countries, notably those that have had a disappointing experience with market liberalization and reform as described in Electricity Market Reform: An International Perspective (2006: Elsevier).

EBL

∗∗ In case of California, the powerful Public Utilities Commission had to publicly admit that it had not done a good job of regulating the industry and proposed to restructure the ESI, substituting market discipline for regulation.

Market Design, Implementation, Performance The previous book had a geographical focus, looking at a number of markets in different parts of the world, describing the reform process and providing a synopsis of how well given markets are working, given their unique characteristics and starting points. This book is focused on an examination of some of the remaining issues, perplexing problems, and persistent questions common to many competitive markets around the world. The issues are grouped into four parts and explored in the following chapters of the book. Some issues, such as industry organization, hybrid markets, or adequacy of investment in infrastructure, are common to nearly all markets. Other topics, such as capacity markets or industry unbundling, may be more pronounced in selected markets with little relevance to others. The organization of the book and a basic overview of the topics are summarized below.

Introduction

13

PART I: Market Reform Evolution The ﬁrst part of the book examines the evolution of market reform from a number of perspectives, including how expectations of market designers, regulators, and policymakers have changed over the past two decades. This change in expectations is due to a more mature understanding of the inherent limitations of the market reform process and a better appreciation of the complexities of the electricity markets as well as the limitations of regulators in achieving what was expected. Chapter 1: Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets Over a century, vertically integrated monopolies gradually evolved under a regulated rate-of-return paradigm. One of the justiﬁcations was the belief that vertical integration results in signiﬁcant economies of scale while making it easier to regulate the players. Among the underlying tenets of market reform is the breakup of vertically integrated companies into sub-components, forcing them to compete when appropriate and removing cross-subsidies that may ﬂow from regulated operations to competitive functions. But two decades after the introduction of market reform, there is clear evidence that many previously unbundled companies have re-bundled, notably by combining generation with retail business. Moreover, there is empirical evidence to suggest that such combinations are more efﬁcient, can manage risks and price volatility better, and may be preferred by investors. At the same time, there is evidence, not universally convincing, that vertical integration – despite its obvious shortcomings – might have offered economies of scale after all. In this chapter, Hung-po Chao, Shmuel Oren, and Robert Wilson argue that the restructuring of the electric power sector is an evolutionary process that must be allowed to develop along a middle path, somewhere between the two extremes of vertical integration and instantaneous unbundling and introduction of competition at both wholesale and retail levels. The authors further suggest that the central issue is whether the physical nature of the industry necessarily implies one extreme or the other. It was long thought that vertical integration of utilities was essential for efﬁcient investments and operations. On the other hand, restructuring has often been motivated by the view that the purported advantages of vertical integration are obsolete and that liberalized markets can work well and bring stronger incentives that are likely to result in more efﬁcient investments and operations. The authors conclude that neither view is conclusive – that pros and cons can be mustered on either side of the argument without any clear indication that one or the other extreme is better. It is suggested that the most important determinants of the optimal degree of vertical integration is how the risks are allocated and managed.

EBL

Chapter 2: Hybrid Electricity Markets: The Problem of Explaining Different Patterns of Restructuring There is general recognition that in many parts of the world, electricity markets have or are evolving into hybrid forms: not completely unbundled and privatized nor fully competitive. Some remain partially privatized or state-owned, some combine generation with supply while others are separated either functionally or physically. In the United States, for example, a relatively vibrant and competitive wholesale market has evolved

14

Competitive Electricity Markets

despite the fact that most consumers continue to receive service under capped, regulated tariffs that do not reﬂect variations in hourly prices in the wholesale market. In this chapter, Aad Correljé and Laurens De Vries point out that while the movement toward liberalized electricity markets around the world is well under way, there is no apparent convergence to a single model or outcome. Textbook prescriptions of integrated markets with decentralized players are not necessarily the norm. In practice, many markets have evolved to a position somewhere between their former, pre-liberalized state and the ultimate state desired or predicted by the markets designers. In some cases, this can be considered a mere transition phase. However, in many markets there does not appear to be even an intention and/or the means of moving toward a fully liberalized state. This chapter examines the evolution of hybrid markets, how they have evolved into their present state, and their implications. Chapter 3: Achieving Electricity Market Integration in Europe Just as the markets and the expectations of policymakers have evolved, so have the role and function of the transmission system operators (TSOs), as they are called in Europe. These new institutions have emerged to be among the most important in the day-to-day operation of the markets by facilitating transactions among various market players. How a given TSO – or ISO in US parlance – is organized and manages given markets has become critically important to how other players – generators, transmission owners, distribution companies, retailers, traders, and investors – interact with one another and how the whole system performs. In this chapter, Nigel Cornwall addresses how the European TSOs have evolved over time and the key role they play in the development of the EC single energy market. The evolution of the European TSOs may be traced to the establishment of the National Grid Company (NGC) in England and Wales and institutional reform in the UK market. The author argues that, in many ways, the United Kingdom’s experiment with the NGC sets the framework and the blueprint for the continental TSOs that followed. While accelerated development has occurred in response to the EC and industry reform programs in recent years, important differences in the regulation, commercial operation, and institutional arrangements in different parts of Europe still remain. This chapter compares and contrasts the evolution of TSOs in six key markets, namely Britain, the Netherlands, France, Spain, Italy, and Germany. The chapter examines legal requirements promulgated by the EC, and looks at differences in the institutional environment between the TSOs. It addresses service provision through both balancing mechanisms and supporting ancillary services, including the commercial arrangements that underpin them. It considers the different regulatory structures and incentives, including compliance and information disclosure and examines drivers for change going forward.

EBL

PART II: Market Performance, Monitoring, and Demand Participation This part of the book examines a number of important market design and performance issues, some peculiar to the United States but others with broad international relevance. Chapter 4: Transmission Markets, Congestion Management, and Investment Current electricity networks are generation-centric, which necessitates a massive transmission and distribution network to transmit the power to major load centers. This served the

Introduction

15

industry well for over a century, especially under a regulated, vertically integrated model. But unbundling of transmission from generation, leaving the latter to private investors, has created a vacuum. There are concerns in some markets, for example, that insufﬁcient investment is going into transmission or that transmission and generation planning have become bifurcated with adverse consequences for both. In the United States, a decade has passed since the introduction of transmission open access. During this period the industry has witnessed signiﬁcant changes in the transmission business with open access under Order 888, the development of Regional Transmission Organizations (RTOs) and, more recently, the evolution of Independent Transmission Companies (ITCs). The period has also seen transmission investment generally lag behind generation investment, putting a greater emphasis on the role of congestion management in market design than has occurred elsewhere. However, there is renewed focus on transmission, with several initiatives to improve investment and transmission markets. In this chapter, Harry Singh reviews the various approaches that have been implemented, data from different markets, and recent initiatives and summarizes some of the lessons learnt. It is fair to say that in nearly all interconnected networks, transmission capacity is scarce during peak demand periods – same as generation capacity – impinging on the ability of market operators to dispatch the least-cost generation options to meet local loads, forcing them to diverge from the so-called unconstrained optimal ﬂow. When transmission is constrained, more expensive local generation must frequently be used to serve local loads, even when lower-cost resources are available somewhere else on the network. This entails signiﬁcant costs, which are ultimately passed on to customers through higher prices. With the increasing recognition of the signiﬁcance of transmission bottlenecks, market operators have devised various methods to make better use of the limited transmission capacity and to encourage both loads and generation to recognize the limitations of the network. Locational or nodal pricing, also called LMP, is among the tools developed in response to this problem. While not everyone embraced this concept initially, some markets in the United States are being redesigned to incorporate it. Concepts such as transmission property rights, transmission auctions, and forward markets to allocate limited transmission capacity have evolved to address these complex issues as well.

EBL

Chapter 5: The Design of US Wholesale Energy and Ancillary Service Auction Markets: Theory and Practice In the United States, the bulk of the electric power system is now under the control of ISOs, which operate multiple markets for energy and ancillary services. Despite some initial difﬁculties in California and elsewhere, and an ongoing need for improvement, these large regional markets encompassing thousands of pricing points are one of the major technical achievements of electricity restructuring in the United States. In addition to providing great potential for economic efﬁciency, much of which is still untapped, these markets will greatly facilitate the growing inclusion of the demand side of the electricity market as well as the absorption of renewable and other supply technologies that are likely to be driven by environmental policy and regulation. In this chapter, Udi Helman, Benjamin Hobbs, and Richard O’Neill discuss the designs of the day-ahead and real-time ISO auction markets for energy and certain ancillary services. The chapter offers a detailed examination of the pricing rules and procedures in these markets. The chapter reviews the history of particular design choices, explaining why they appeared more desirable than alternative proposals. Differences among the ISO

16

Competitive Electricity Markets

designs are also examined, with a focus on PJM and New York ISO. It also provides a numerical example that illustrates primary market features, such as day-ahead and realtime locational marginal pricing (of energy, congestion, and losses), unit commitment decisions, virtual bidding, reliability commitments, revenue sufﬁciency guarantee, ﬁnancial settlements, and disposal of auction surplus payments that accrue when congestion and losses are priced at the margin. The example can be easily replicated, allowing for its use as a teaching and study tool.

Chapter 6: The Cost of Anarchy in Self-Commitment-Based Electricity Markets A persistent design issue in the context of electricity markets is whether plant unit commitment decisions should be made centrally by the system operator or by individual generators. Although a centrally committed market can, in theory, determine the most efﬁcient commitment, they have been shown to suffer some equity and incentive problems. A self-committed market can overcome some of these incentive issues, but will generally suffer efﬁciency losses from not properly coordinating commitment and dispatch decisions between individual generators. In this chapter, Ramteen Sioshansi, Shmuel Oren, and Richard O’Neil examine the issue of dispatch efﬁciency raised by the design of markets based on central versus selfcommitment by determining a set of “competitive benchmarks” for the two market designs. Comparing the total commitment and dispatch costs of the two markets provides a bound on the productive efﬁciency losses of a self-committed market.

EBL

Chapter 7: Market Power and Market Monitoring Deregulation is good public policy only to the extent it furthers competition. While proper market design is important, experience to date suggests that the success of competitive electricity markets depend ﬁrst and foremost on structural conditions that do not distort market outcomes. One of the important conditions is that structural market power must be addressed as an integral part of market reform. Several states in the United States have required or encouraged the divestiture of generation by the former vertically integrated electric utilities. Even then, as the California crisis of 2000–01 has shown, those markets could experience local market power in load pockets when constrained transmission imposes limits on power imports. Market monitoring units have a crucial role to play in detecting and preventing such market manipulations. In this chapter, Parviz Adib and David Hurlbut discuss the market power problems encountered in those markets. The authors make an important distinction between scarcity pricing, which is essential for attracting investments in competitive markets, and artiﬁcially induced high prices resulting from the exercise of market power. They describe the tools that are most useful to market monitors to detect market power abuses and mitigate or minimize their impacts. To carry out their monitoring and investigative functions fairly and effectively, market monitors must be independent from the market participants and the system operator, and their ﬁndings and conclusions must not be inﬂuenced by political considerations. The authors conclude that the burden is on legislative and regulatory authorities to ensure that market monitors are empowered to do their job effectively, that their independence is respected, and that their ﬁndings and recommendations are taken into consideration.

Introduction

17

Chapter 8: Demand Participation in Restructured Markets Demand elasticity is a much talked about topic, but demand has not been effectively integrated with supply-side resources. Most markets in the United States and internationally are currently supply-focused, in the sense that hourly demand is taken as a given, and generation is adjusted to meet the variable demand. Demand inelasticity is a major problem in most markets, notably those with needle-sharp peaks and low reserve margins. But making demand more elastic remains an elusive goal, technically as well as institutionally. In this chapter, Jay Zarnikau reviews the challenges associated with facilitating demandside participation in competitive markets, provides a summary of various demand-side initiatives in restructured markets in North America, and contributes a case study describing some of the problems faced in promoting demand-side participation in the Texas market, often cited as the most successful retail market in the United States.

PART III: Capacity, Resource Adequacy, and Investment The third part of the book is focused on interrelated issues of capacity, resource adequacy – or lack thereof – and infrastructure investment. These issues, while germane to all markets, are nevertheless treated differently in different parts of the world.

Chapter 9: Resource Adequacy: Alternative Perspectives and Divergent Paths

EBL

In electricity markets characterized as energy-only, generators are rewarded for the energy delivered to the network when and if they are dispatched. In such energy-only markets, little if any reward is paid to generators for having extra capacity available for long-term system reliability. Although most markets also operate ancillary service (AS) markets, where generators are paid relatively small amounts for providing spinning reserves, reactive power, and the like, these are rewards to generators for providing an operational safety margin in real time – not for long-term system reliability. The issue of whether capacity markets, where generators are paid signiﬁcant amounts for capacity – regardless of whether they are dispatched or not – continues to be important for at least two reasons: First, in countries where there is signiﬁcant amount of hydropower,12 thermal generators are likely to be infrequently dispatched in years when ample hydro capacity is available. This creates problems during drought periods because lack of sufﬁcient incentives to invest in thermal generation may result in severe shortages during such episodes. 13 • Second, in countries where excess reserves are externally mandated, as in the United States, average wholesale prices tend to be depressed, resulting in inadequate investment in future capacity. Further complicating the problem is the presence of low offer caps in many markets, which has the effect of limiting scarcity revenues to generators •

12

Examples include Chile and Colombia, both of which have capacity charges and are considered problematic. 13 The preliminary evidence from the Nordic market where a severe drought was experienced without major meltdown (Amundsen et al., 2006), however, suggests that even hydro-dominated systems can operate without a capacity market.

Competitive Electricity Markets

18

when supplies are short. But designing a capacity payment scheme that encourages adequate investment in new capacity without introducing serious unintended side effects has proven to be a monumental challenge. The question is this context is twofold: •

Do we need to pay for capacity above and beyond energy to ensure adequate supply reliability? • If so, how can we design incentives that produce the desired beneﬁts without serious adverse consequences? There is a growing body of literature on this topic but the fundamental debate has not been settled,14 at least not in the United States, which is the focus of the chapter by Parviz Adib, Eric Schubert, and Shmuel Oren. The authors review the controversial evolution of resource adequacy mechanisms in the United States while examining alternative approaches considered in the Texas market. The necessary conditions and a potential transition mechanism to a sustainable energy-only approach are also discussed. Chapter 10: The Evolution of PJM’s Capacity Market Continuing in the same vein, in this chapter Joseph Bowring describes the evolution of capacity markets in PJM since their inception in 1999 in response to regulatory requests to facilitate retail competition through modiﬁcations and eventually wholesale changes in 2006. The chapter examines the theory underlying the need for capacity markets, the ways in which PJM capacity markets have evolved and functioned, and how reality has matched or not matched with theory. The chapter also addresses the interactions between energy markets and capacity markets and evaluates the joint functioning of the two markets based on empirical results. Total and net revenues from these markets are presented as well as the effectiveness of local price signals in incentivizing investment. The chapter provides a critical self-examination of PJM’s new capacity market, the Reliability Pricing Model (RPM) in light of the actual reliability performance of the PJM markets. The discussion provides an example of how a speciﬁc capacity market has been designed and how well or poorly it has been functioning in the single largest centrally dispatched market in the world. The experience of PJM – and the reasons it has chosen to introduce a capacity market – are relevant to discussion of capacity markets elsewhere.

EBL

Chapter 11: Resource Adequacy and Efﬁcient Infrastructure Investment Resource adequacy and investment continue to be a major concern in a number of countries, as described in Electricity Market Reform: An International Perspective (2006: Elsevier). The jury on how well markets can or will provide the necessary funds to feed the industry’s voracious appetite for long-term capital is still out. This problem is not unique to developed countries; many developing countries are attracted to market reform primarily because they need massive amounts of foreign investment in their power sector. 14 Paradoxically, markets in PJM, New York, and New England have decided to proceed with capacity payment schemes while ERCOT and MISO have examined and rejected them, and California is still debating if such a mechanism is needed.

Introduction

19

Table 3. Resource adequacy and reliability Attribute

Objective

Resources

Incentives

Resource adequacy

Have enough installed capacity to fulﬁll demand in the medium and long term

Installed capacity subject to availability and technical constraints

Have price signals that provide the incentives to invest in capacity, recovering operation and maintenance costs

Reliability

Have the resources to respond to short-term ﬂuctuations

Provision of ancillary services to satisfy minimum reliability standards deﬁned by the authority

Have price signals that explicitly account for ancillary services provision

Source: Adopted from R. Raineri, 2006, personal communications.

The main concern is that insufﬁcient investment may be going into generation, distribution, and – most notably – the transmission sector. The evidence is mixed and it is not entirely clear if markets are failing to deliver because of delays in processing the price signals, regulatory uncertainties, or other impediments. And there are questions about how best to mitigate such market shortcomings, if indeed they exist and are real.15 This issue is closely linked to capacity market, but there are clear distinctions between adequacy and reliability as highlighted in Table 3. Concerns about supply inadequacy resulting from the “missing money” problem described in Cramton and Stoft (2006) have led to proposals for creation of dual markets where generators get paid for energy on the one hand and capacity on the other. In this chapter, Alan Moran and Ben Skinner argue that a reasonably efﬁcient market has been achieved in Australia without regulation of generation. The outcome, which is not without some fragilities, has generally been due to less government intervention than seen in some other markets, with a higher reserve intervention price, less distortive consumer price caps, and a genuine level of retail competition that provides good market signals for new capacity.

EBL

PART IV: Market Design Issues This part of the book is devoted to several important issues that are not necessarily caused by but have become more pronounced since the introduction of market reform. Incorporation of renewable energy technologies, the role of distributed generation, and the possible menace of global climate change are not market reform issues per se. Addressing them, however, has become more urgent and important with the introduction of market reform. Each topic arguably deserves an entire book, but has been squeezed into a single chapter. 15

There are those [e.g., Moran (2006)] who attribute any shortcomings of the private sector to continued meddling and regulatory interference.

20

Competitive Electricity Markets

Chapter 15 uniﬁes many of the concepts covered in preceding chapters of this volume, explaining how Brazil has attempted to put an order into its vast electric power sector following the introduction of market reform. Chapter 12: Promoting Electricity from Renewable Energy Sources – Lessons learned from the EU, the United States, and Japan Advocates of renewable energy technologies believe that governments must support and subsidize them, especially if one is concerned about global climate change. Among the arguments made is that renewables must gradually replace dwindling supplies of ﬁnite fossil fuels, ultimately replacing them. As the costs of fossil fuels rise, renewables will become more cost-effective, and subsidies can be reduced and eventually removed. While such arguments are gaining growing support, there is little agreement on how best to subsidize the renewable energy industry – in ways that do not interfere with the efﬁcient workings of competitive markets. Currently, a variety of support and subsidy schemes are in place in different parts of the world with no particular rhyme or reason for some of the targets or subsidies provided. In the United States, for example, a number of states have adopted renewable portfolio standards (RPSs), with varying targets and target dates in the absence of a cohesive national policy16 (Fig. 6). Similar issues confront policymakers in Europe where very signiﬁcant amounts of renewable energy have been developed in the recent past in response to highly generous

EBL

Fig. 6. Renewable portfolio standards in effect in various states in the United States. 16

The deadline for California is 2010 (not 2017 as indicated in Figure 6) and there is an Executive Order issued by California Gov. for a 33% by 2020. At the federal level, there have been proposals to impose a 15% national target.

Introduction

21

subsidies. Aside from the form and amount of subsidies, the European renewable energy now generates a large amount of power, which must be absorbed by the existing network. This is of major concern in the European context, because renewable power occasionally plays havoc with the transmission grid as well as with conventional generation, which must adjust its output to absorb the intermittent load. More research is needed to determine how various schemes, for example, renewable portfolio standards, favored in the United States, compare to feed-in tariffs, favored in Europe. In this chapter, Reinhard Haas, Niels Meyer, Anne Held, Dominique Finon, Arthuro Lorenzoni, Ryan Wiser, and Ken–ichiro Nishio explore the evolution of renewable energy resources (RES) in Europe, the United States, and Japan over the past 20 years. The authors examine a variety of programs in different countries describing their success and failures. European policymakers are currently discussing the application of harmonized policies throughout Europe. Chapter 13: Distributed Generation and the Regulation of Electricity Networks Current electricity systems are designed based on large centralized generation and topdown transmission and distribution networks serving the loads. In recent years this paradigm has been challenged as more decentralized power plants are located close to load centers and connected to the distribution network with advantages that include, apart from the economic and environmental advantages, the security of supply. Distributed generation (DG), including combined heat and power (CHP) and renewables, enjoys high policy priority in many countries and is likely to have a bright future. Yet, it does not always easily ﬁt into today’s centralized power systems. In this chapter, Dierk Bauknecht and Gert Brunekreeft examine how a growing share of DG can be integrated into liberalized electricity markets from a regulatory point of view. While the technical debate on DG integration is well under way, the question as to how the governance structure of the electricity system can best promote and integrate DG has received less attention and has mainly been focused on the design of support mechanisms for individual plants. This chapter focuses on electricity networks and the role of network regulation for the integration of DG. It is often argued that DG can reduce network costs. However, in many cases networks need to adapt to DG, which may entail incurring additional costs. Even if the overall beneﬁts of DG are positive, additional network costs represent a disincentive for network operators to connect DG. The regulatory framework within which they operate is crucial for efﬁcient integration of DG. This includes unbundling rules and the regulation of connection charges and network tariffs.

EBL

Chapter 14: Global Climate Change and the Electric Power Industry Economists view global climate change as an externality. So long as there is no price or penalty on emitting CO2 into the atmosphere, we emit more than we should. The same is true for disposal of industrial waste or SOx . The most efﬁcient way to control CO2 emissions would be to price it – or put a mandatory cap on what can be emitted – which will have the same result. Once CO2 emissions are priced, ﬁrms will take measures to limit their emissions using the most cost-effective options at their disposal. Firms that cannot meet their quota will pay others, buying credits. In this chapter, Andrew Ford begins with the premise that the accumulation of greenhouse gas (GHG) emissions in the atmosphere is a serious problem that warrants serious

22

Competitive Electricity Markets

action and provides a summary of the initiatives at the international, national, and regional levels. The chapter focuses on emission trading scheme (ETS) to limit GHG and the central role of the electric power sector in combating global climate change.

Chapter 15: Reform of the Reforms in Brazil: Problems and Solutions This chapter starts with a discussion of the “reform of the reforms,” namely the fact that most markets require secondary and tertiary meddling to ﬁx the problems introduced by the initial process of market reform. No market in the world has been entirely immune to this, although some have required minor adjustments and ﬁne-tuning as opposed to major post-operative surgery. Nowhere has this problem been more pronounced than in Brazil. In the Brazilian electricity supply industry, the ﬁrst market-oriented reform ran into investment troubles. The reform of the reforms, led by the Lula administration, purports to overcome this through a combination of free bilateral contracts between free consumers, traders, distributors, producers, and regulated contracts between generators and distributors through auctions. In this chapter, Lizardo de Araújo, Agnes Maria de Aragão da Costa, Tiago B. Correia, and Elbia Melo analyze the characteristics of the Brazilian electric power sector in terms of original market reforms as well as the ongoing process of “reform of the reforms,” covering power auctions, investments in generation and transmission, and efforts to regulate electricity prices. In many ways, this chapter uniﬁes many of the concepts described in earlier chapters of the volume, serving as a case study.

EBL

References Amundsen, E., von der Fehr, N.H., and Bergman, L. (2006). The Nordic market: Robust by design? In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Araujo, Jo˜ao Lizardo R. Hermes de (2006). The case of Brazil: Reform by trial and error? In Electricity Market Reform: An International Perspective. (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Cramton, P. and Stoft, S. (2006). The convergence of market designs for adequate generating capacity. Prepared for California Electricity Oversight Board, 25 June. Dyner, I., Arango, S., and Larsen, E.R. (2006). Understanding the Argentinea and colombian electricity markets. In Electricity Market Reform: An International Perspective. (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. The Economist (2006). 11 February. Federal Energy Regulatory Commission (2006). Report to Congress on competition in the wholesale & retail markets for electric energy. Washington, D.C., 5 June. Goto, M. and Yajima, M. (2006). A new stage of electricity liberalization in Japan: Issues and expectations. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Haas, R., Glachant, J.M., Keseric, N., and Perez, Y. (2006). Competition in the Continental European electricity market: Despair of work in progress? In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Joskow, P. (2006a). In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Joskow, P. (2006b). Markets for power in the United States: An interim assessment. The Energ. J., 27(1). Littlechild, S. (2006). Foreword: the market versus regulation. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier.

Introduction

23

Moran, A. (2006). The electricity industry in Australia: Problems along the way to a national electricity market. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Newbery, D. (2006). Electricity liberalization in Britain and the evolution of market design. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. O’Neill, R., Helman, U., Hobbs, B., and Baldick, R. (2006). Independent system operators in the United States: History, lessons learned, and prospects. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Raineri, R. (2006). Chile: Where it all started. In Electricity Market Reform: An International Perspective. (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Sioshansi, F. and Pfaffenberger, W. (2006). Why Restructure Electricity Markets? In Electricity Market Reform: An International Perspective. (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Sweeney, J. (2006). California electricity restructuring, the crisis, and its aftermath. In Electricity Market Reform: An International Perspective. (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Treblicock Michael & Hrab, Roy (2006). Electricity restructuring in Canada. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier. Tschamler, Taff (2006). Competitive retail power markets and default service. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Elsevier.

EBL

EBL

This page intentionally left blank

Part I Market Reform Evolution

EBL

EBL

This page intentionally left blank

Chapter 1 Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets HUNG-PO CHAO1 , SHMUEL OREN2 , AND ROBERT WILSON3 1

ISO New England, USA; 2 University of California, Berkeley, USA; 3 Stanford Business School, USA

Summary

EBL

This chapter critically reviews the argument for vertical integration in the electricity industry, and also the argument for restructuring based on unbundling of its products and organizations in favor of market mechanisms. The authors conclude that both arguments are deﬁcient, and that a balanced mixture of vertical integration and liberalized markets is superior to the extremes. Their central conclusion is that efﬁcient management of the risks inherent in the electricity industry requires that restructuring retain universal service for the core of non-industrial customers who rely on regulated rates smoothed over time to recover the costs of service. 1.1. Introduction This chapter addresses basic economic issues posed by restructuring. The central issue is whether the overall technology of the industry – wholesale generation, transmission, and retail service – necessarily implies more or less vertical integration. It was long thought and is still being argued by many that vertical integration of retail utilities was essential for efﬁcient investments and operations (e.g., see Michaels, 2006). On the other hand, restructuring has often been motivated by the view that the purported advantages of vertical integration are obsolete, that liberalized markets can work well, and that they bring stronger incentives that are likely to result in more efﬁcient investments and operations (e.g., California Public Utilities Commission, 1993). The argument presented here is that neither view is conclusive – that pros and cons can be mustered on either side without any clear indication that one or the other extreme is better. In prior work (Chao et al., 2006) the authors argued that restructuring of the electricity industry should develop along a middle path between the extremes of vertical integration and liberalization of wholesale and retail markets. This middle path establishes the boundaries of the ﬁrm – the extent to which a retail utility should retain some degree of vertical 27

28

Competitive Electricity Markets

integration. A key element of this choice is the make-or-buy decision about whether to own and manage supply resources, or to rely on wholesale markets via either spot purchases or longer-term contracts. A middle path also requires restructuring of regulatory policies and redeﬁnition of the regulatory compact to recognize the effects of investment, purchasing, and contracting decisions by utilities in the context of liberalized wholesale markets, and to strengthen incentives for efﬁcient operations and demand response. Moreover, the optimal extent of vertical integration is ultimately determined by the requirements for efﬁcient allocation of risk bearing. After restructuring, the most important determinants of the optimal degree of vertical integration concern risk management, which affects the cost of capital – the ultimate measure of ﬁnancial risk – and supply reliability and resource adequacy – the ultimate measures of physical risk. (See also Correljé and De Vries, Chapter 2 in this volume.) From the perspective of risk management, the mutual interests of suppliers of generation and retail service enable risk sharing that mitigates ﬁnancial risks. Depending on local circumstances, their shared interests imply a greater or lesser degree of reliance on markets and contracts, or on direct ownership that perpetuates some degree of vertical integration. For example, a utility might meet some resource adequacy requirements by contracts or by purchases in capacity markets, and also own generation facilities that serve its core retail customers within a regulatory scheme that continues the traditional regulatory compact, albeit with stronger incentives from market forces and performance-based regulation. Section 1.2. begins by reviewing the case for vertical integration of utilities that prevailed through most of the twentieth century. Section 1.3. examines anew these arguments in the current context and ﬁnds them greatly altered – in part by the evident successes of some aspects of restructuring. The discussion of economic issues in Section 1.2. includes a summary of explanations of vertical integration in the literature. This discussion is necessary because ideas from this debate have greatly inﬂuenced restructuring decisions by regulators and legislators, especially in Europe recently. It also clariﬁes the distinction between ﬁnancial and organizational “unbundling” of a utility’s vertical components – wholesale generation, transmission, and retail service – and unbundling of the corresponding products. In the regulated era, the organization of the electricity industry stemmed from vertical integration of utilities in all respects, while in the past decade much reorganization aimed at segmenting utilities into their vertical components in conjunction with unbundling of their products. In several cases, organizational unbundling of ﬁrms was seen as a necessary or desirable complement to unbundling of products to facilitate liberalized wholesale markets. Although organizational disintegration was rejected in most other liberalized industries (transport, telecommunications, etc.), regulators and legislators favored dissolution of vertical organization in the electricity industry for reasons that are reviewed. Section 1.4. reviews some of the unsolved problems of liberalized markets, including both those that cannot be solved efﬁciently by market processes and those that have not yet been solved adequately by market restructuring. Section 1.5. develops the case that risk management considerations are major determinants of the degree of vertical integration in terms of organization and ownership and vertical contracting. Section 1.6. concludes by outlining some implications for the evolution of restructuring. This discussion introduces scenarios in which a desirable degree of vertical integration coexists within liberalized wholesale markets for unbundled products, and which allow a utility to serve core customers at regulated rates while others opt to purchase from competing suppliers.

EBL

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

29

1.2. The Historical Motives for Vertical Integration The origin of vertical integration in the electricity industry lies in a dominant public interest. Like other infrastructure industries – water, transport, communications – the energy industries were recognized as essential for economic development. Universal service, efﬁciently supplied at minimum cost, was imperative. In many countries these needs in the case of electricity were addressed by monopolies conducted or owned by local or national governments, and in some cases by government projects or subsidies; e.g., in the United States by the Tennessee Valley Authority, Bonneville Power Administration, Western Area Power Authorities, and the Rural Electriﬁcation Administration. The prevalence of government monopolies and government-sanctioned monopolies had three sources. One was technical, resulting from the advantages of alternating current synchronized over grids spanning large regions. Another was economic, resulting from the large scale of transmission and distribution (T&D) systems and the large scale of some generation facilities, especially hydroelectric dams but also the most efﬁcient coal-ﬁred and, later, nuclear plants. The third was ﬁnancial, because the government was the sole or chief source or guarantor of sufﬁcient capital at low cost. All three reﬂected the capital intensity of the technology used by the power industry, certainly in T&D, and in combination with fuel intensity in the case of generation. Historians of economic development view the twentieth century, in part, as an era of accumulation in which massive investments established the infrastructure on which a modern economy depends. (Chandler, 1969; Devine, 1983) The industry’s organization differed in those countries like the United States, Japan, and Germany that relied heavily on investor-owned utilities (IOUs). Although Nebraska and some municipalities developed public power systems, and federal projects were important elsewhere, within the United States most major urban areas depended on IOUs for provision of retail service. The role of IOUs stemmed from a conjunction of public and private interests. The public interest in universal service at minimum cost was matched by ﬁrms’ interest in obtaining ample capital at low cost. The states established Public Utility Commissions (PUCs) to regulate the industry (except federal regulation of interstate trade), with authority to mandate the quality, conditions, and terms of retail service (Bonbright, 1961). In return, each utility obtained an exclusive regional franchise, except for municipal utilities and rural cooperatives, which were exempt. In principle, this was a retail monopoly but it evolved into a total franchise that encompassed local supply, transmission, and distribution as well as retail service. A state’s grant of monopoly franchises on transmission and generation was artiﬁcial since it derived from comprehensive cost-of-service regulation rather than basic economic considerations. It was fundamentally at variance with federal legislation and regulation, but enforced by each state’s control of siting of facilities, cost recovery from retail rates, and authority to exclude independent power producers (IPPs) from selling to retail customers. Under the old “regulatory compact,” risk management was provided through an insurance mechanism by vertical integration along the electricity supply chain. The single utility ownership of generation and transmission facilities buffered wholesale price volatility. Retail regulation smoothed the rate effects of cost changes on customers, imposed an obligation to serve, and offered utility shareholders a reasonable opportunity of recovering investments with a largely assured rate of return. Although all the risks – both physical and ﬁnancial – were socialized to a high degree, customers bore the residual risk. Importantly, a utility was assured full recovery of prudently incurred investments and expenses, including the cost of capital. This part of the regulatory compact was implemented by nearly level retail rates; that is, a utility’s recovery of an approved cost (one accepted into the rate base) was amortized over many years, with repayments obtained

EBL

Competitive Electricity Markets

30

from retail revenues. The regulatory compact was a perfect means of obtaining capital from private sources to build a growing industry – because cost recovery was assured, utilities obtained capital from ﬁnancial markets at low cost without drawing on public funds. Amortization of cost recovery reduced risks for lenders and shareholders, and equally, it reduced the volatility of rates paid by retail customers. For regulators, cost-of-service regulation brought difﬁculties judging prudency and measuring costs, and they were often dismayed by a utility’s weakened incentives for cost minimization and strengthened incentives for capital-intensive projects (CPUC, 1993; Joskow, 1997). But until the last decade before restructuring these deﬁciencies were viewed as of second-order importance compared to the advantages. A utility’s monopoly on local generation and T&D was implemented by vertical integration of all aspects, including organization. The electricity industry has a linear supply chain from fuel to generation to transmission to distribution to service delivery. Each utility integrated backward from retail service to encompass at least generation, and occasionally some fuel sources. There were two motives for extension of a utility’s monopoly backward into the supply chain, and with it the resulting vertical integration. One was the advantage of a single coherent investment strategy. Given the load-duration proﬁle and the costs of building and operating generators, there is a particular mix of generation technologies that serves the load at least the overall cost in the long run. There is also an optimal conﬁguration of the transmission grid and locations of generators, and moreover, an optimal substitution between local generation and transmission to access distant generation – as well as occasional use of local generation to alleviate congestion on transmission elements, sustain voltage, etc. The second motive was the advantage of consolidated operations. Centralized dispatch of generation and transmission had the explicit objective of minimizing the total cost of serving the load subject to constraints intended to ensure service reliability and protect the transmission grid from cascading failures.

EBL

1.2.1. Theoretical framework These motives were always based on ideal realization of the alleged advantages in investments and operations. In reality the actual results were often driven by practical ﬁnancial considerations, as explained below. Even so, a substantial body of economic theory was constructed to explain the prevalence of vertically integrated utilities (e.g., Williamson, 1975, 1985). Its main ingredients were as given below: •

Public good. The T&D system is the enabling infrastructure of the power industry. Tight control, operating on very short time frames, is required to sustain service reliability and to avert cascading failures of grid elements and generation units. Also necessary are uniform standards and procedures among interconnecting segments of the grid. • Natural monopoly. Duplication of T&D facilities is wasteful except where it improves grid security or service reliability. • Economies of scale. Natural monopoly was extended to generation by citing the large size and capital requirements of efﬁciently scaled units and plants. This argument applied mainly to hydro projects and base-load plants using coal and later nuclear fuels. • Economies of scope. This catch-all category (in principle, a subset of economies of scale) cites advantages from tight coordination, such as the above-cited advantages

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

31

of centralized investment and operations. It also includes advantages from substitution (e.g., generation capacity usable for either energy production or a contingency reserve, and generation used to alleviate transmission congestion), and the possibility that standards, technology, information systems, and skills used for one kind of generation are applicable to other kinds and to engineering control of the grid. Savings in metering, billing, and ﬁnancial settlements are sometimes included in this category. • Economies of transaction costs. Despite its name, this category refers not to costs of metering and billing, but to difﬁculties and risks in contracting. Its premises include asset speciﬁcity and incompleteness of contracts. A seller’s investment in a transmission or generation facility is irreversible and long lived, and the facility cannot be moved or used for another purpose. The value of the investment is therefore tied speciﬁcally to expected use by or sale of output to buyers. If there is a single buyer then an initial contract between them might seem to ensure that the seller obtains the value he anticipates when he commits to investment and construction. But a contract that covers all contingencies is usually infeasible, and in those unlikely contingencies that are not covered (or if the buyer can renege) the seller might not be able to renegotiate with the buyer to recover the sunk costs of investment. Anticipating this, the seller might not undertake the investment initially. This scenario is the basis for the argument that contracts may be insufﬁcient to stimulate adequate investments, and therefore vertical integration of the seller and the buyer might be necessary to ensure that efﬁcient investments are undertaken. These technical explanations of vertical integration did not, however, address the more practical aspects that were constantly at the forefront of regulatory considerations. These were dominated by ﬁnancial considerations that are described next.

EBL

1.2.2. Financial motives In keeping with their primary responsibilities, PUCs focused on capturing the advantages of vertical integration for retail service. Cost-of-service regulation was the means to obtain mandated universal service at minimum cost and with high reliability. In some ways a vertically integrated utility was easier to monitor, its total costs were easier to measure, and it could be held directly accountable for deﬁciencies of quality or reliability. Universal service required subsidies to those residential and commercial customers who were more expensive to serve, and vertical integration offered the expedient of relying on implicit cross-subsidies rather than explicit ﬁnancial subsidies. Although industrial customers were especially disadvantaged by this policy, the inefﬁciencies of cross-subsidies were secondary to the political inﬂuence of residential and commercial customers. The magnitude of cross-subsidies declined in later years as the development of efﬁcient plants of small size and the growth of co-generation enabled an industrial customer to negotiate lower rates, since it had the option to self-generate to serve its own load. Most important for PUCs was that retail rates could be smoothed over time by amortizing the utility’s recovery of its costs, and the cost of capital could be minimized by indirectly invoking the credit of the state. Full recovery of costs via retail rates necessarily implies that retail customers ultimately bear nearly all of the ﬁnancial risks; indeed, this is the ﬁrst fundamental principle of the regulatory compact. For many customers their aversion to volatile rates is profound, in part because on short timescales they have limited options to alter usage patterns or to invest in alternative appliances and production technologies, and generally they cannot obtain ﬁnancial hedges against ﬂuctuating rates.

32

Competitive Electricity Markets

However, the second fundamental principle is that cost recovery is amortized so that rates can be smoothed inter-temporally over long periods. The feasibility of this scheme stems basically from the difference between the high volatility of fuel and power prices in the short term and their low volatility over the long term. Short-term volatility is mainly cyclical and load-based, ranging from daily variation to seasonal weather cycles to business cycles. Thus, retail customers are exposed only to secular risks and trends, and only gradually. These include trends in fuel prices and generation technologies, and inevitable mistakes such as misestimates of the amount and location of load growth. However, this rosy scenario was upset in the years after the oil embargo of 1973 and before restructuring in the 1990s by cost overruns for nuclear plants and by the prices guaranteed to “qualifying facilities” (QFs) as speciﬁed in the Public Utilities Regulatory Policies Act (PURPA) of 1978. But until well into the 1980s cost-of-service regulation was generally viewed as successful in spite of the difﬁculties during the 1970s and 1980s from gyrating fuel prices, monetary inﬂation and high interest rates, and technical advances that rendered major investments inefﬁcient. The natural monopoly aspects of T&D systems implied regulation and control of rates. Most investments in T&D facilities could not be recovered by marginal-cost pricing or by congestion pricing. Therefore, cost-of-service regulation that provided recovery of investment and maintenance costs extended naturally to T&D. In later years there were instances of performance-based regulation (Hunt, 2002), and, rarely, of merchant transmission investments in direct current lines, but overall the expansion of the grid was a massive investment in infrastructure that continued until in the United States and Canada it is now composed of only two interconnected systems plus one within Texas. Although essentially a public asset, the grid is largely privately owned by utilities and ﬁnanced mainly by recovering the costs from charges included in the rates paid by retail customers. Like some other infrastructure networks (railroads, telecommunications, gas pipelines) it was regulated according to principles of contract carriage – until superseded in 1996 by common carriage when the Federal Energy Regulatory Commission (FERC) required open access and nondiscriminatory pricing. Reliance on private ownership and contracting prevailed in some countries (e.g., Germany and Japan) while state-owned transmission companies developed the grid in others (e.g., the United Kingdom, France, New Zealand, and Scandinavia). The latter developed systematically but those relying on local utilityowned transmission developed through increasing interconnections among them as energy trading increased and the utilities increasingly relied on exchange agreements to improve reliability. The ﬁnancial aspects of generation were fundamental motives for vertical integration. Mentioned previously were the role of capital intensity, the scale economies of base-load plants before smaller gas-ﬁred plants were developed in the 1980s, and possible economies of scope. These are reinforced by the great variation of loads on short time frames and the resulting high volatility of prices in spot markets, plus longer-term secular trends. Since a generation plant has a lifetime of 20 to 40 years, its inherent value is largely unaffected by short-term price volatility. Moreover, the supplier and a buyer such as a utility have mutual interests to ensure each other against price variations, since every price that is good for one is bad for other. Thus, one might surmise that long-term contracting will ensure investments in generation capacity; indeed, the investor can use the contract as security to obtain loans to ﬁnance construction. This scenario is jeopardized, however, by two factors. One is that a 20- or 40-year contract differs from utility ownership of a plant only in its lack of direct investment and operational management and control, but requires comparable justiﬁcation to the

EBL

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

33

PUC that it is prudent, and further that it is invulnerable to the supplier’s default or bankruptcy. The other factor is that the investor or its lender prefers a contract that ﬁxes both price and quantity, while the utility prefers ﬂexible dispatch to meet changing loads and overall demand growth, so their mutual interest in price insurance is diminished by their opposing interests in “volumetric” insurance. All the intermediate contract forms (e.g., option contracts, tolling contracts) require at least one and usually both parties to bear risks of one kind or another. For the investor or its lender, risk reduces the value of the investment, and for the utility, any residual risk borne by its shareholders is inferior to assured cost recovery if it undertakes the investment itself – in the usual circumstance that it can obtain capital at lower cost than non-utility investors – and thereby transfers the risk to ratepayers. These ﬁnancial considerations in generation investments are variants of the factors invoked in the analysis of economies of transaction costs described earlier. A contract sufﬁciently complete to deal with all contingencies is too complex to be practical, and even if it were feasible and could miraculously insure both parties so that for the utility it is not inferior to inclusion of the new plant in the rate base over the lifetime of the plant, approval of the contract in a prudency review would be problematic – and in some contingencies might not be enforceable. For example, if the contract extended over decades then during a prolonged period of low prices the supplier might default on the contract.1 The realistic contracts are therefore short-term (usually a few years, rarely 10) and incomplete, with both parties bearing shares of the price and volumetric risks. Until shortly before restructuring, this picture doomed non-utility generation from the start. With its low cost of capital and other advantages, a utility could always undertake a more efﬁcient portfolio of investments than could private investors exposed to price and/or volumetric risks, and do so without sacriﬁcing dispatch control. In the United States, PURPA ﬁrst forced utilities to purchase generation from the QFs. This wedge initiated further opening of wholesale markets. Next were the Energy Policy Act of 1992 and the FERC’s ensuing Orders 886 and 888 that forced utilities to allow open access to unused transmission at nondiscriminatory prices. This wedge also initiated bilateral contracting between non-utility power generators and large industrial customers in the late 1990s.

EBL

1.2.3. The hidden assumptions The justiﬁcations of vertically integrated utilities contain hidden assumptions. As described in Section 1.3., many of these were revealed by actual experience after restructuring. Several that bear on how one interprets the foregoing arguments for vertical integration are listed below. •

A utility obtains capital at lower cost than its supplier. This assumption might seem to contradict theories of ﬁnance, but in fact it was realistic before restructuring. The seeming contradiction stems from the fact that one can invest equally in shares of the contracting parties (the independent generator and the utility). It therefore seems that an investor can hedge against price and volumetric risks that affect the seller and buyer oppositely, and therefore the seller and buyer should obtain

1 This assumes omission of provisions in recent contracts that enable the utility to take over the plant immediately in the event of default. Since restructuring, utilities’ contracts with suppliers have increasingly included stringent provisions to protect against the consequences of default.

34

Competitive Electricity Markets capital at comparable costs. In fact, however, cost-of-service regulation transfers the utility’s risks to ratepayers, and an investor cannot easily invest in shares of the general ratepayer population, especially over the extended period of cost recovery. If the transfer were not mixed with cross-subsidies then an investor might hedge by buying shares of a generator and an industrial customer, but this strategy was ineffective in the regulated era. Thus, one must view utilities’ lower cost of capital as both a cause and a consequence of vertical integration under cost-of-service regulation. • The utility is the sole buyer of generation. Those arguments that invoke economies of transaction costs depend heavily on the bilateral character of supply contracts and on the dominant monopsony power of the local utility, derived from its exclusive retail and transmission franchises. The advent of bilateral trading of bulk power supplies over wide regions, enabled by open access to transmission on nondiscriminatory terms, undercut this argument, and it evaporated with the arrival of multilateral trading in power exchanges managed by system operators. Both parties to contracts now have comparable options outside their bilateral relationship, namely to sell or buy at prevailing spot prices or to contract on a longer-term basis with competing alternatives. The prospect that a generation investment would be hostage to the sole utility buyer of energy, and to its transmission system, must now be seen as a consequence of the utility’s exclusive franchises, and thus an artifact of regulatory policy. • The owner of the transmission grid has the authority to manage it. Another artifact of regulatory policy was a utility’s right to exclude access to other parties, even if ample capacity were available after serving its native load. Arguments were mustered that the utility needed exclusive control to assure reliability and security, but with the advent of regional system operators (in 1998 in the United States for those systems not previously organized as power pools, but years earlier elsewhere) this argument dissolved. Indeed, the efﬁciency gains from regional dispatch and transmission allocation are now widely recognized. The argument for economies of scope also dissolved since nearly all components of this category now pertain to the system operator. What remains is the vestigial argument that coordinated planning of transmission expansion and generation investments might be more efﬁcient if it were centralized, but even so the advantages occur at the level of regional systems rather than utility service territories. • Incentive effects are minor compared to economies of scale and scope. The engineering expertise of the utilities was always admired, and the PUCs took some credit for the ﬁne performance by attributing it to the high standards they imposed (one day of outage in 10 years was the prevailing standard), as well as the advantages of vertical integration in establishing operating standards and procedures and in internalizing the public-good aspects of system security. This credit was less convincing after engineering operations and many skilled personnel were transferred to the FERC-regulated regional system operator. The regulated era was always beset by complaints that the commercial parts of utilities were complacent bureaucracies, and motivated more by cost recovery at inﬂated costs of capital than by minimizing overall costs (e.g., by favoring capital-intensive projects).

EBL

In later years these complaints led PUCs to experiment with performance-based regulation, rate caps, negotiated rates for industrial customers, and other devices. Behind these complaints, however, lay the basic fact that a utility’s ﬁnancial incentive was muted by assured cost recovery and thus by full insurance against contingencies – and after the PUC

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

35

accepted an asset into the rate base, by insurance against errors of judgment. The source of this comprehensive insurance was the regulatory compact. In some states, guaranteed cost recovery escalated rates to levels higher than neighboring states that then attracted away industrial and commercial ﬁrms seeking lower energy costs, or required negotiated industrial rates to forestall self-generation. In the United States restructuring was precipitated by one such state when the California PUC announced in 1993 that it would consider new regulatory principles and policies based on greater reliance on markets (see Borenstein et al., 2002). A chief consideration was the view that generation investments would be more efﬁcient if private investors rather than ratepayers were to bear the consequences of erroneous judgments. Since insurance mutes incentives, one of the alternative policies it outlined, the one adopted in 1994, withdrew some or all of the provisions for assured recovery of investments in generation. These four assumptions, mostly hidden in the standard justiﬁcations and explanations of vertical integration, are typical of a longer list. They are emphasized here because they exemplify a tendency for the merits to be addressed within the regulatory policies, institutional structure, and market rules that sustain vertical integration. Section 1.3. takes the opposite approach and describes the fundamental changes that preceded and followed restructuring. Restructuring introduced new regulatory policies and market structures that reﬂected a new view that vertical integration is not intrinsic to the electricity industry; indeed, one can assemble a comparable argument that electricity is amenable to an industrial organization that relies heavily on liberalized markets.

EBL

1.3. The Case Now for Liberalized Markets

The case for liberalized wholesale markets is now examined from several perspectives. First we provide a brief review of the situation after restructuring. Then theoretical arguments are examined in light of empirical evidence. In both cases the discussion includes some developments in the two decades before restructuring in the United States, along with the experience after restructuring. The analysis focuses on those aspects that indicate the future role of vertically integrated utilities. The discussion is organized around four utility functions after restructuring: system operations, wholesale markets, retail service, and generation.

1.3.1. System operations after restructuring Some aspects of the industry remain unaffected by restructuring. The importance of universal service is reinforced now because economic development depends on technologies that rely on reliable power supplies. The traditional role of lighting is now supplemented by digital information and communication systems. Heating and cooling applications that were considered secondary are now considered essential for much commercial activity. The increased role of reliability enhances the public-good character of the transmission system, which remains a natural monopoly. But management of the grid is now viewed as a technical task, one that the engineering profession is well able to conduct, and that can meet the highest standards when its span is regional. This consensus developed early in those systems with national transmission companies, but in the United States it evolved from cooperative power pools and from the observation that local utilities were well able to integrate supplies purchased from QFs into their routine operations and dispatch procedures.

36

Competitive Electricity Markets

Control areas conﬁned to utility service territories are an impediment to coordination over the wide areas now required. The magnitude of coordination problems at seams will eventually reveal the optimal span of regional operations, but only the largest utilities are viable candidates for retaining their own control areas. Equally, a utility’s motive to hoard its transmission facilities to serve its native load impairs efﬁcient allocation of generation and transmission capacity. Recognizing this, orders by American regulators have steadily mandated or encouraged formation of independent system operators (ISOs) and regional transmission organizations (RTOs) with authority to manage regional transmission systems on a daily basis, and responsibility for ensuring open access and nondiscriminatory pricing. These changes are partly organizational, but also ﬁnancial since they invoke principles of common carriage, remove “pancaking” of transmission charges, and impose charges to recover costs of ancillary services and re-dispatch to alleviate transmission congestion. 1.3.1.1. Consolidation of grid and market control From an economic viewpoint, the organizational specialization represented by system operators is notable for two features. Most important is that the externalities inherent in grid operations are handled by engineering procedures that enforce standards for reliability and security. The most obvious externality stems from the grid’s role as a public good enabling transmission among locations. Because the grid is vulnerable to cascading failures, automatic switches open lines and disconnect generators to minimize damage to facilities. The consequences for suppliers and retail customers are partly transitory due to loss of power production and consumption, but their equipment and appliances can also be injured, and industrial customers can lose goods in intermediate stages of production and incur the costs of idle labor. A subtler externality stems from differences between the technologies of supply and demand. On the supply side, sufﬁcient fast-response reserves must be available to meet most contingencies because the ramp rates of generators are limited. The high value of ramp rate on the supply side has no counterpart on the demand side, since customers care only about whether power is on or off.2 If there is only one customer (e.g., a utility) then it sees clearly that continuous power availability depends on its provision and payment for reserves. But if there are many customers then each knows that what it pays for reserves has little effect on its own access to power. This is a classic free-rider problem (if there are many customers then each prefers that others provide and pay for reliability) since the marginal effect of any one customer’s contribution to reserves has a small effect on overall system reliability and therefore a small effect on the reliability of that customer’s supply. The free-rider problem is signiﬁcant even when the customers are a few utilities, since each sees that its own marginal value from a marginal dollar expended for reserves depends heavily on how much others contribute.

EBL

2 A mathematical model leading to the conclusion that a completely decentralized market for energy and reserve capacity does not obtain full efﬁciency is by Chen et al. (2004). They conclude: “The decentralized market tends to depress ancillary service prices, which leads to the failure of the second welfare theorem. At each time the two market prices represent the market value of the on-line capacity of ancillary and primary services. From the system operator’s viewpoint, the ancillary service is more valuable because of its higher ramping rate. However, from the utility’s viewpoint, both services are identical, as long as they are available. In the decentralized market, the utility does not consider the ramping rate, since this is a constraint on production rather than consumption. As a result, the utility does not want to pay a higher price for the ancillary service.”

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

37

Thus, without regulatory intervention and engineering command-and-control, markets for reserves are bound to be inefﬁcient or even to collapse. These considerations extend beyond daily operating reserves to the general problem of ensuring adequate supply resources, including both generation and transmission capacity. To an investor, constructing a plant that will be idle most of the time seems a waste because it is called only rarely to meet contingencies, and this is equally true of a transmission line constructed to provide a backup for others that might fail. Thus, some form of payment to idle capacity is necessary to ensure that investments are sufﬁcient. The practical issues of implementing resource adequacy requirements are addressed in Chao et al. (2006). In principle, each customer could be charged for the rate of variation in its load, and thus pay for reserve generators with high ramp rates. However, a basic economic advantage of a regional electricity system is that variations of customers’ loads largely cancel out, and aggregate loads are substantially predictable – the day-ahead prediction of the aggregate peak load in an hour is usually considered to be accurate within 3% or 4%. For this reason, retail pricing has largely ignored the possibility of charging for load variation other than on the basis of a customer’s load-duration proﬁle over an extended period such as a year and to some extent, by real-time energy prices revised every few minutes. One could charge for load variation in some conditions of aggregate variation, such as the morning ramp at the beginning of the workday, or for increasing use of air conditioning when temperatures escalate on a hot summer afternoon, but in fact metering and pricing have not been developed to this degree of reﬁnement.3 Instead, system operations provide a buffer to ensure the steady matching of supply and demand, re-dispatching online generators and calling on reserves as necessary to follow the aggregate load. The buffer is partly automatic, since generators equipped with governors and automatic controls (AGC) adjust power output in response to frequency variations detected by sensors. The automatic buffer provides an operator with an interval, usually considered to be about 10 minutes, in which to re-dispatch and call on reserves, beginning with hydro and spinning reserves, and then non-spinning reserves that require start-up and synchronization. The basic economic signiﬁcance of system operations is that they supplant price-mediated market mechanisms in favor of command-and-control to ensure reliability and real-time matching of supply and demand. Market processes might conceivably be used to balance supply and demand almost continuously, but the costs and risks are too extreme to make them feasible on the short time frame that is relevant in a power system. Thus, the economies of scope invoked to justify vertically integrated utilities in the United States are now mostly obtained by consolidation of grid management and wholesale spot markets in system operators. These developments in the United States imitated earlier initiatives in other countries with national transmission companies. Linking of the Scandinavian state-owned systems into the coordinated multinational NordPool was one model, and the other was the England–Wales Pool that began in 1989 [e.g. Hunt (2002) and its analogs in Argentina, Alberta, and Australia – followed later by New Zealand, Spain, and others, e.g. EPRI 2002 and Rudnick et al. (2005) and Barker et al. (1997)]. Some consolidated systems encountered initial operational and economic problems, and all were reformed later in some ways, but the basic principle that regional systems are more efﬁcient and more reliable has not been challenged.

EBL

3 However, after the California crisis the state provided $35 million to fund installation of interval meters at all large industrial and commercial customers, accounting for about a third of the aggregate load. Estimates of the costs of meters and the resulting beneﬁts typically imply substantial net beneﬁts; (cf. Borenstein 2004).

38

Competitive Electricity Markets

1.3.1.2. Unbundling of products A second feature of system operations is unbundling of wholesale supply into constituent products such as energy, reserves, and transmission, which nevertheless are allocated jointly by a consolidated spot market conducted by the system operator. Unbundling of supplies into relatively homogeneous products like energy, reserves, and transmission recognizes the basic scarce resources managed by engineers. These products must be differentiated by attributes like time and location, and operational constraints like a generator’s start-up time, minimum energy output, and ramp rate. Earlier arguments that this complex mix of products and attributes could not be efﬁciently priced are now, after successful implementations of time- and location-differentiated nodal pricing, conﬁned to the operational constraints that involve nonconvexities, such as the start-up costs of a generator. Similarly, the various categories of regulation and operating reserves (spin, non-spin, replacement) are now priced systematically by recognizing that speed of response is the scarce resource, and therefore prices for slower reserves are limited by the feasibility of substitution with faster reserves. Unbundling wholesale supply into standard products enables efﬁcient allocation among multiple parties, and more speciﬁcally, unbundling facilitates markets and settlement procedures for multilateral trade. Markets for medium-term bilateral contracts for energy (e.g., Eltermin in NordPool and the UK Power Exchange) also rely on standard product speciﬁcations.4 The responsibilities of a system operator now exemplify the economies of scope argument, since the engineers protect reliability and system security while also facilitating and/or conducting wholesale markets, and, indeed, procure resources needed for grid management from these same markets. For instance, the real-time balancing market is a market for both buyers and sellers, and also provides the bids from which the operating engineers obtain resources to follow the load, alleviate transmission congestion, and sustain voltage. Even so, spot markets have been organized quite differently among various system operators. Some reﬂect technology, as in the ﬁrst version of NordPool where ample hydro resources allowed self-scheduling and an emphasis on energy trading, and zonal pricing sufﬁced since the chief transmission bottleneck was between Norway and Sweden. Those adopting the England–Wales Pool model focused on efﬁcient day-ahead scheduling of thermal generators using a comprehensive optimization of dispatch. The ISOs in the US northeast also adopted this model since it simply extended the procedures of pre-existing power pools. Their relative success during and after the California crisis in 2000–01 motivated FERC to propose a Standard Market Design (SMD) that is now adopted also by California, among others (FERC, 2002). The SMD’s emphasis on comprehensive day-ahead optimization of all aspects does more than consolidate spot markets for products like energy, reserves, and transmission, since the optimization allocates available generation and transmission capacity by including unit commitments and scheduling along with assignments to energy generation and reserve status. Settlements are based on hourly locational prices for energy, reserves, and transmission, but the scarce resources are capacities rather than ﬂows.

EBL

4

A peculiarity of bilateral contracting is that the resulting demands for transmission need not result in an efﬁcient allocation. That is, with bilateral contracting the use value of transmission depends on the pairings of sellers and buyers, whereas multilateral trading can provide the maximum value obtainable among all pairings. This potential problem has been insigniﬁcant in systems that have enough multilateral trading (e.g., 30–40% is scheduled centrally in PJM) to ensure that transmission is accurately priced to reﬂect scarcity values at the margin.

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

39

California considered the pool model but opted for an initial design closer to NordPool, including self-scheduling of generation and loads, a power exchange separate from the ISO, zonal pricing of transmission, and simple market clearing for standard products rather than comprehensive optimization (cf. Sweeney, 2006). This design was plagued by loose coordination and gaming of market rules almost from inception, and then virtually collapsed in a series of events initiated by scarcity of imported supplies from hydro sources.5 In the United States it is now widely accepted that the system operator must retain tight control, and in particular, that its authority must extend beyond reliability to maintenance of orderly wholesale markets. For instance, FERC now allows ISOs to impose various protective measures on generators: must-offer obligations, bid caps, automatic procedures for mitigation of market power, obligations to respond to dispatch instructions, mandatory scheduling of maintenance, penalties for large deviations from day-ahead schedules, and a dozen more interventions that might be listed. Analogous problems occurred elsewhere (e.g., New Zealand) when supplies were scarce, but there is no exact parallel to the California crisis. The panoply of ad hoc protective measures now imposed in the United States have not been adopted widely because other countries provide system operators with ample authority to ensure reliability and more discretion in managing their markets. The United States is unique in requiring an ISO to adhere rigorously to the terms of its FERC-approved tariff and market rules, and avoid any inﬂuence on energy markets. The California ISO was allowed the least discretion and was least able to bring its markets under control, but inability to weather a crisis is inherent in the strictures placed on all ISOs. The exact opposite can be seen in the United Kingdom’s reformed New Electricity Trading Arrangements (NETA) system, where the transmission system is owned and operated by an independent transmission company (ITC), the National Grid Company (NGC), rather than a non-proﬁt bureaucratic ISO. NGC must adhere to a Balancing Code for settlements but, within a scheme of performance-based regulation that rewards reductions in its grid management charge, it has wide discretion to manage the grid, including taking positions in the energy market to acquire reserves and counter market power. The NETA system is also the opposite of FERC’s SMD, since all energy trading is conducted through private power exchanges for bilateral contracts, and physical feasibility is established hours-ahead rather than day-ahead (Newbery, 2006).

EBL

1.3.1.3. Lessons learned The lessons learned from the recent experience with system operations can be summarized thus: The organization, governance, and procedures of a regional system operator are very important and very complicated. Creating such an entity is a major task from an engineering viewpoint, and the design of its markets is equally challenging from an economic viewpoint. The complexity of system operations ensures that an initial design must be revised as deﬁciencies are discovered. The task is worthwhile because the regional scope can enhance reliability and improve overall efﬁciency of the short-term allocation of generation and transmission capacity. The vigor and growth of wholesale energy markets attests to gains from trade from system operations and markets on a regional scale. On the other hand, the California crisis was a salutary warning that wholesale power markets are 5 The origins and history of the California crisis are described by Blumstein et al. (2002) and Wilson (2002). An empirical analysis of the role of market power during the crisis is by Borenstein et al. (2002).

40

Competitive Electricity Markets

fragile when supplies are scarce, demand is inelastic, and utilities are obligated to serve but ﬁnancially exposed. It is necessary to address these vulnerabilities – in terms of both resource adequacy and ﬁnancial exposure – as argued later, but a backup remedy should be adequate authority and discretion for the system operator to intervene to stabilize its markets. That is, it should not rely solely on engineering procedures and established market rules when active intervention can stabilize markets or suppress the inﬂuence on market prices of dominant suppliers, and thus its scope should include some provisions for active management of markets. There is no indication yet that any one design is best. The diversity of designs now working reasonably well in various countries and within the United States suggests that local factors are important. A principal reason that local considerations can be determinative is that engineering management of a transmission system is so well developed (and virtually uniform worldwide) that it assures most of the gains from a regional system regardless of which among several alternative market designs are used. Wholesale markets for energy might be bilateral or multilateral, decentralized or optimized, provided system operators have adequate means to ensure reliability and allocation efﬁciency. A relevant comparison observes that the United Kingdom relies on private markets for bilateral contracts, Australia relies on an energy-only real-time market, and NordPool uses a day-ahead and real-time market; and zonal pricing of transmission congestion sufﬁces in the latter two. Seemingly quite different are the day-ahead and real-time markets in the Pennsylvania–New Jersey–Maryland (PJM), New York, and New England systems that include unit commitment and scheduling and co-optimization of energy and reserves, and insist on the importance of nodal pricing of congestion. Yet differences in performance among these systems must be considered of second-order importance compared to their overall successes. This in particular requires that markets are workably competitive; that is, incentives promote productive and allocative efﬁciency. The California experience especially demonstrates the need for a design that discourages gaming of market rules. Gaming is essentially always due to a market imperfection, usually an unpriced scarce resource, and therefore a signal that efﬁciency can be improved by the measures that also eliminate gaming. For instance, the costly “dec game” in California was possible because congestion charges were imposed only between large zones and only day-ahead, which enabled those who caused intra-zonal congestion to escape congestion charges day-ahead and then be paid in real-time for alleviating the congestion they caused. The fact that nodal pricing eliminates the dec game illustrates the more general principle that even though many market designs are possible it is still true that efﬁciency requires that all scarce resources are priced, in this case intra-zonal transmission capacity.

EBL

1.3.2. Wholesale markets after restructuring This subsection outlines the economic argument that a liberalized wholesale power market is potentially an efﬁcient means of allocation among buyers and sellers of energy. This argument assumes, of course, that engineering aspects are conducted by a system operator whose procedures supplant market processes on the short time frames relevant for protecting reliability and ensuring continuous matching of supply and demand. The discussion focuses ﬁrst on the spot market for multilateral trading, then extends the analysis to the forward market for bilateral contracting, and then examines the incentives for efﬁcient investments. The argument is mainly theoretical but mentions indications that its

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

41

predictions are conﬁrmed in some markets now in operation. The complexity of the markets now conducted by ISOs is evidence that implementation is difﬁcult, but the concern here is whether wholesale markets are efﬁcient in principle. 1.3.2.1. Models of restructuring The pace and scope of liberalized wholesale markets differ greatly among countries. The focus below is on those with comprehensive markets, some of which extended their scope gradually (e.g., NordPool, and Australia, which evolved from Victoria’s VicPool) while others liberalized in a single decisive act (e.g., England–Wales in 1989). There are two basic models: In one model the utility remains the single buyer, but regulators require the utility’s “make or buy” decision to consider competing offers from IPPs. In the United States this approach took an extreme form due to the 1978 PURPA that essentially required a utility to pay its avoided cost for supplies from small plants (less than 80 MW) that used co-generation of heat and power or renewable sources of energy. The second model allows a market for bilateral contracting between IPPs and large industrial customers, augmented by provisions for enhanced opportunities for trading among utilities. This model was in effect in the United States in the period after federal regulators required that utilities provide open access to transmission on nondiscriminatory terms, and it too stimulated substantial investments by IPPs. It is currently seen in the interim phase of the European Union’s directives for partial liberalization, and more speciﬁcally in the organization of the electricity industry in Germany. It may be that most of the gains and fewer problems are obtained with these intermediate forms of liberalization. Indeed, even in the United States those restructured systems that have allowed utilities to remain substantially integrated are cited as more successful – and the complete divestiture of gas- and oil-ﬁred plants by California’s utilities is cited as one source of the crisis there. However, the aim here is to examine the viability of fully liberalized wholesale markets, and therefore to focus on those systems with comprehensive markets. The argument for restructuring depends crucially on its most important innovation, which is management of the regional transmission system by a system operator such as an ISO (O’Neill et al., 2006). Initially the discussion simply assumes that a system operator manages the transmission grid, although some of the difﬁculties encountered by a system operator are mentioned in passing. This enables bypassing the public-good and naturalmonopoly aspects of transmission, and some of the operational tasks required to assure reliability and system security. At the end of this subsection are comments on deﬁciencies due to inefﬁcient allocation of risks, and an outline of the role of regulated utilities in improving risk management. The main motive for a market is to realize gains from trade. This motive originates in some separation of ownership. Power markets (and cooperative power pools) began with exchange agreements among utilities to enhance reliability, extended to long-term trading of energy supplies based on differing costs and asynchronous loads, and then expanded to daily “economy” trading to minimize generation costs. The gains from trade in a modern market stem from its regional scope and from substantial separation of ownership between sellers (generators) and buyers (utilities and other load-serving entities, LSEs). Separation of ownership between generators and LSEs may reﬂect advantages from specialization, but restructuring has proceeded more on the premise that the main advantages come from stronger incentives. One incentive effect is supposed to be more efﬁcient investments when investors in generation bear the consequences of their decisions and operating decisions, rather than

EBL

42

Competitive Electricity Markets

relying on utilities’ assured recovery of costs. The second incentive effect is vigorous competition among generators when they are sufﬁciently numerous and small to have little inﬂuence on prices. Competition among generators is imperfect in most wholesale markets, and therefore typically requires some regulatory interventions; e.g., price or bid caps and must-offer obligations are typical. The vigor of competition depends ultimately on sufﬁcient investments in generation and transmission capacities. At a minimum this requires that no ﬁrm is “pivotal” in the sense that its capacity is needed to meet a peak load in a transmission-constrained area. Over a longer time frame, competition stems from contestability, in the sense that incumbents cannot maintain high prices without stimulating new investments by entrants. Contestability has become a more effective constraint as smaller combined-cycle units can be installed in a few years, and for peak loads and ofﬂine fast-response reserves, combustion turbines (CTs) can be installed in a few months. Competition among LSEs is usually considered to be of secondary importance because their service obligations and the price inelasticity of retail demands limit their opportunities for strategic behavior to arbitrage between forward and spot markets. 1.3.2.2. Spot markets The spot markets conducted by ISOs are multilateral; that is, they allocate supplies offered by several sellers to several buyers. They are also “smart markets” in that energy trades are optimized subject to constraints on transmission capacities and generators’ ramp rates, required quality attributes (frequency, voltage), and operational procedures that ensure sufﬁcient reserves to meet contingencies and avert cascading failures of equipment. A multilateral market is feasible only if several basic requirements are met. These include standardized commodities and qualities, accurate metering, explicit market rules, and settlement procedures that include assured creditworthiness of market participants. These requirements and other enabling aspects are now routine among system operators. Participants must subscribe to a contractual agreement that imposes reciprocal obligations, such as compliance with dispatch instructions, scheduling of outages for maintenance. After allegations of manipulations during the California crisis in 2000–01, the importance of a rigorous code of conduct and steady scrutiny by an independent market monitor is now universally accepted. Here it is taken as given that the mainly smooth operations of spot markets conducted by system operators are evidence that the basic enabling requirements of spot markets are feasible, and their implementation is now well developed. The gains from trade obtained from an ISO’s multilateral spot market are reduced in proportion to the extent of bilateral contracting in forward markets. Even so, gains remain because forward contracts account incompletely for contingencies and imperfect predictions of loads. Even Britain’s NETA system, which relies on forward contracting up to a few hours ahead of real-time operations, conducts a real-time balancing market in which operators re-allocate supplies to follow the load, alleviate congestion, and procure supplementary reserves. Most other ISOs rely on a day-ahead market as the primary multilateral market for energy trading because it can be integrated with unit commitments and scheduling, and engineering operations can ensure physical feasibility in advance by establishing reserve assignments and alleviating transmission congestion. In the United States, FERC insists that physical feasibility is established day-ahead so that the realtime balancing market is less volatile and less vulnerable to gaming that might threaten reliability or cause extreme prices. But the choice between the extremes represented by the NETA system and the tight controls enforced in the United States evidently depends on local circumstances; e.g., the greater prevalence of transmission congestion and tighter energy supplies in some regions may account for the choice in the United States.

EBL

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

43

The efﬁciency of a multilateral spot market depends ultimately on whether pricemediated transactions are sufﬁcient. They are insufﬁcient if efﬁciency depends on signiﬁcant public goods or other externalities. One can interpret the system operator’s management of the grid as ensuring the public-good aspects of reliability. The other main externality is environmental, and in many countries it is addressed by markets for emission allowances. Even so, the more fundamental impediment to efﬁciency concerns the scope of wholesale markets. In principle, efﬁciency requires that each scarce resource has its appropriate price. To a great extent this requirement has been addressed by multiple simultaneous markets for energy, reserves, and transmission, and further, by prices for each that are differentiated by time and location; i.e. by spatially differentiated “nodal” prices established at short intervals. But a peculiarity of power markets is that there are other resources that occasionally are scarce; e.g., reactive power for local voltage support. Also, some products are imperfect versions of the resource that is actually scarce; e.g., the reserve categories reﬂect imperfectly the scarcity of fast-response resources, and in particular the key quality attribute, which is ramp rate or start-up time. Schemes have been tried to establish prices for reactive power and ramp rate, but system operators usually ﬁnd it sufﬁcient to rely on engineering procedures and standard reserve categories. In the occasional instances that its markets do not provide adequate resources to manage the grid, engineers retain authority to issue dispatch directives that are settled according to rules for “out of market” transactions. A prevalent deﬁciency in the United States is persistent under-scheduling: when day-ahead generation schedules derived from the energy and reserve markets provide insufﬁcient online generation to meet predicted loads, the ISO must make additional unit commitments and pay generators whatever portions of their start-up costs are not recovered from market sales. Some ISOs encourage arbitrage between day-ahead and real-time prices (by allowing “virtual” bids day-ahead that are not backed by physical resources or loads) in an attempt to reduce under-scheduling. In sum, one can conclude that from an operational viewpoint the market conducted by a system operator is inherently inferior to fully consolidated operations within a vertically integrated utility, but increasingly the prevalent view is that the disparity is small, and that such markets are mainly successful. The sufﬁciency of price-mediated transactions must also be considered from the viewpoint of market participants. At a mundane level, the obvious burdens that the ISO imposes on participants (bidding, responses to dispatch directives, settlements, etc.) are likely greater than those in vertically integrated systems, and in some the ISO’s expenses and therefore its grid management charges (called “uplift”) are higher than anticipated. Britain’s NETA system is notable for using performance-based regulation of NGC that rewards reduction of this charge. A basic deﬁciency of simple market clearing is that it cannot cope directly with thermal generators’ nonconvex cost components such as startup and no-load running costs, and nonconvex operating constraints such as minimum generation rates and maximum ramp rates. Since this deﬁciency arises partly from the definition of the traded products, alternative deﬁnitions have been proposed (e.g., enabling a base-load generator to bid to supply energy steadily over the day) but not widely adopted. The two main alternatives are self-scheduling of generators by their owners (e.g., NordPool, NETA, California, Texas) and in those systems that inherited the operating procedures of power pools (e.g., PJM, New England, New York), central optimization of schedules for those units not committed to bilateral contracts (using “three part bids” that include the ﬁxed-cost components and generators’ reports of their operating constraints). Unit commitments by the ISO require that the portion of ﬁxed costs not recovered from market prices is uplifted. The ISO schedules additional units to enhance reliability, but in

EBL

44

Competitive Electricity Markets

several systems it is alleged to depress energy and reserve prices. New England introduced separate semi-annual markets for ofﬂine fast-start reserves mainly to provide adequate revenues to attract sufﬁcient installations of CTs needed to protect against large contingencies. As described later, a general issue is whether an ISO’s markets stimulate investments in an optimal mix of generation technologies. 1.3.2.3. Forward markets Forward markets for longer-term bilateral contracts enable both parties to hedge against price and/or quantity risks. Bilateral contracting plays a large role in all wholesale power markets due to the high volatility of spot prices. Some contracts impose physical requirements but most are essentially ﬁnancial hedges against spot prices, as for example in a “contract for differences” (CFD) in which the seller and buyer insure each other against deviations of the spot price from the strike price speciﬁed in the contract. The central role of forward contracting was evident in the California crisis when the California utilities, which were largely prohibited from contracting forward, encountered severe ﬁnancial difﬁculties while other utilities in nearby states in western United States that faced equally high spot prices were not jeopardized because they relied on spot markets for small shares of their procurements, usually less than 10%. When the state intervened to stabilize the California market its principal tactic was to secure long-term contracts that thereafter provided the bulk of the utilities’ requirements. It is important to realize, however, that long-term contracts are risky in a different way. In California the state signed contracts with generators that speciﬁed prices that turned out to be exorbitant in the long run; moreover, the contracts speciﬁed ﬁxed quantities that in some circumstances were excessive. This debacle repeated California’s earlier mistake in the 1980s when it offered QFs long-term contracts at prices that later were revealed as excessive. Although contracts written as options could have avoided these unfavorable outcomes, this experience illustrates the more basic source of the risks inherent in longterm contracts. Wholesale power markets are inherently vulnerable to systemic risks, i.e., risks that cannot be fully dissipated by mutual insurance between contracting parties. Systems with substantial hydro resources are vulnerable to prolonged droughts, those with mainly thermal plants are vulnerable to changing fuel prices and new-generation technologies, and on the demand side, both are affected by seasonal and annual weather patterns, business cycles, and other large-scale economic developments – some cyclical and some reﬂecting secular trends. In the regulated era, systemic risks were moderated by recovering costs from retail rates that varied slowly over extended periods. Restructuring introduced a new tension between the advantages of forward contracting in insuring against short-term spot-price volatility and the risk that the strike price or promised quantity speciﬁed in a contract would turn out ex post facto to be unfavorable to one or the other party. Trading of standard intermediate-term bilateral contracts is vigorous in those systems (e.g., NordPool and Britain) with power exchanges that effectively minimize search and transaction costs. These contracts are usually for ﬁxed quantities, which is somewhat anomalous since on general grounds one might expect other forms, such as option contracts, to be useful hedges against quantity risks. However, recent years have brought a greater variety of contract forms and tolling agreements, and some of the innovative contract forms that might be used to ensure resource adequacy. In all systems the major share of power generation is covered by forward contracts; e.g., in PJM bilateral contracts account for about twice the volume traded in its spot market, and in Britain’s NETA system they are nearly 100%. FERC’s SMD supposes that the bulk

EBL

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

45

of power trading will rely on forward contracts. In particular, it aims to conﬁne the spot market to adjustments day-ahead and real-time to address contingencies and to assure physical feasibility and reliability. The volume of real-time trading is usually under 10% in well-functioning systems – except during California’s crisis when real-time trading approached 50%, presenting dire threats to reliability. Since then, forward contracting in California has been essentially mandatory, enforced with penalties for deviations from day-ahead schedules that exceed 5%, until recent changes to its market rules. Regulators usually exempt a utility from prudency reviews for a moderate amount of purchases via intermediate-term bilateral contracts, provided they are standard contracts traded in organized markets with adequate competition and transparency. But this exemption does not apply to long-term contracts and to any transactions that hint of self-dealing with afﬁliated generation companies. During the initial phase of restructuring, divestiture of a utility’s generation assets was eased in those systems that included so-called vesting contracts in terms of the spin-off of the generation subsidiary or sales of its assets. The assets were bundled together with contracts that ﬁxed the prices and quantities of continued sales to the utility for several years. This procedure avoided self-dealing while providing the requisite ﬁnancial hedges for the seller and buyer after divestiture. It also had a profound effect on dominant suppliers’ inﬂuence on prices in spot markets. In general, a generation ﬁrm’s gain from withholding supply or bidding higher to raise spot prices is reduced in proportion to the amount of its capacity that is committed to ﬁlling the requirements of forward contracts. The market inﬂuence of dominant suppliers in the England–Wales system increased after the expiration of vesting contracts, and analogous effects are now evident in Australia. Regulatory policy has therefore often focused on measures to ensure that both generators and utilities are substantially hedged against spotprice volatility by forward contracts. A signiﬁcant impediment, however, is that utilities subject to competition from other LSEs and from IPPs are reluctant to sign very long contracts in view of the risk that their service obligations might change substantially. Their role as the retail provider of last resort (POLR) exacerbates this risk; e.g., an industrial customer might opt for bilateral contracting with an IPP when market prices are low and then later opt to return to service from the utility when prices rise. Many generation companies insist on the importance of long-term forward contracting. Their incentive to insure against spot-price volatility is greatly strengthened by effects on their costs of capital. The proﬁtability of an investment in a new plant depends crucially on the cost of capital obtained from lenders and equity investors. Long-term contracts for major portions of the plant’s capacity reduce the risk of the investment, and thereby the rate of return demanded by sources of funds; indeed, a lender often treats long-term contracts rather than the physical asset as the main security for a loan.

EBL

1.3.3. Retail service after restructuring Cost-of-service regulation was long implemented in a way that contradicted its premise. A utility was reimbursed for costs incurred, but not until late in the regulated era was a customer charged the actual incremental cost of the service provided. Equally inefﬁcient was the absence of differentiated service conditions that would allow customers a range of choices beyond simply the number of kilowatt-hours (kWh) to draw from the system at the standard cents-per-kWh (¢/kWh) price. Limited choice and uniform pricing were expedient in the years when the infrastructure of the electricity industry was being established, universal service was a dominant consideration, the technology of retail service

46

Competitive Electricity Markets

delivery was primitive, and metering and billing were major impediments to service differentiation. Nevertheless, regulators continued these policies long after the limitations that previously justiﬁed them had relaxed. Cross-subsidization from industrial and commercial to residential customers was a basic feature of universal service, but there were others derived from technical and economic factors. The most obvious considerations stemmed from the public-good aspect of grid security and the fact that service reliability is largely uniform. A customer’s service might be interrupted or curtailed, but quality attributes like frequency, waveform, and voltage are inherently uniform in a system with alternating current. Interruptions and curtailments can be imposed selectively only with costly metering and control technologies, or by direct communication that was practically conﬁned to large industrial customers. The uniformity of quality attributes implied a basic tension between those customers who preferred lower rates for lower-quality and less reliable service (e.g., heating and cooling applications) and those who preferred higher rates for higher-quality and more reliable service (e.g., lighting and industrial production). This tension was resolved in favor of greater reliability for several reasons. One was the technical advantage of a highly secure grid and the low cost of extending high quality to all customers, and another was the importance of high quality in promoting economic development. But the basic enabling feature was cross-subsidization that in effect charged premium rates to industrial customers for high quality that allowed lower rates for residential customers and extension of universal service. There was always an array of special provisions (e.g., low offpeak rates for street lighting and other municipal services) and special considerations (e.g., provision of the highest quality to hospitals and other essential facilities) but here we focus on the main tensions among industrial, commercial, and residential customers. These customer categories are used here as surrogates for the much more complex diversity of preferences among customers; and even for a single customer, diverse preferences in relation to different appliances and technologies (e.g., heating/cooling, lighting, production, information/communication). This heterogeneity implies efﬁciency gains from service differentiation, but in the electricity industry there were, and remain, severe technical and cost barriers that preclude full differentiation of service conditions and rates. One can describe restructuring as a late stage of the more general trend to improve overall efﬁciency. At the retail level, this entailed unbundling of service components, pricing based on incremental cost, and, inevitably, a declining role for cross-subsidization. This trend included all the infrastructure industries, but the discussion here addresses only its effects in the electricity industry and the developments in the retail sector, and the next subsection addresses developments in the generation sector. Both cases emphasize the efﬁciency improvements that were sought through restructuring, and largely ignore the political resistance that inevitably accompanied the elimination of subsidies. It also ignores the role of subsidies from the government in some developing countries.6

EBL

6 The major industrial countries decided long ago that the electricity industry must cover its costs from retail rates paid by customers. A major exception occurred when the state of California issued debt to fund a 10% retail rate reduction during the ﬁrst years of restructuring, and then during the crisis assumed ﬁnancial responsibility for the utilities’ wholesale procurements, and later issued debt to fund it. National transmission companies indirectly rely on the government’s good credit but costs are recovered from customers.

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

47

Economic theory predicts that a uniform ﬂat rate causes inefﬁciencies for three reasons, all due to the heterogeneity of customers. •

A uniform rate does not reﬂect directly the incremental costs of services demanded by different customers in response to that rate; customers are more or less sensitive to the costs they impose on the system. • A uniform rate applies to an undifferentiated commodity, whereas customers have differing preferences for the various quality attributes of service. • A ﬂat rate foregoes opportunities to recover infrastructure costs with less distortion of incentives. Economists invoke two general theories about how to recover ﬁxed costs from regulated retail rates in a way that promotes overall efﬁciency. One theory is due to Ramsey and its application to pricing by utilities is due to Boiteux and Mirrlees (summarized in Wilson, 1993). Generally, it supposes that the utility would run a deﬁcit if infrastructure costs were not recovered by commodity taxes included in retail rates. Its characteristic implication is that customers with more inelastic demands should pay larger shares of the deﬁcit. This theory was largely rejected by regulators due to the characteristic feature of retail service that demands are most inelastic among residential customers and least among industrial customers. The other theory aims to sensitize customers to the cost implications of their demands for service. This theory has been applied in two very different ways, depending on whether costs are measured in the long run or the short run.

EBL

•

A prominent application of pricing that reﬂects long-run costs is used in France, where tariffs are designed to emphasize the long-run implications for the utility’s investments. A commercial or industrial customer pays a “demand charge” that depends on its peak load and then, for each kilowatt within that peak load, an energy charge that depends on the number of hours that kilowatt is used during the year (which requires a special meter).7 In effect, this scheme charges the customer for its load-duration proﬁle over the year. • Rates that reﬂect short-run costs are differentiated by time or events. The simplest tariffs distinguish only between peak and offpeak periods. Equally simple in concept is real-time pricing based on the actual system marginal cost in each hour, but metering costs have conﬁned applications to large industrial customers. Until recently, the costs of communication or control, metering, and billing were long the impediments to differentiation of rates based on peak–offpeak periods or real-time pricing (Zarnikau, Chapter 8 in this volume). It remains one reason that rates differentiated by times or contingencies rarely extend to residential and small commercial customers, but equally important are customers’ own costs of real-time communication and control, and most fundamentally, their reluctance to bear short-term price volatility and inability to obtain ﬁnancial insurance. 7

This form of retail pricing, called a Wright tariff in the United States, was used in the early years of the industry. See Wilson (1993, section 2) for an extended description of the tariffs in France, circa 1990.

48

Competitive Electricity Markets

1.3.4. Generation operations and investments after restructuring Most systems allow self-scheduling of generation committed to bilateral contracts, including generation within utilities, and some allow self-scheduling for all generation. In the United States those systems that give the ISO responsibility for all other scheduling impose substantial requirements for adherence to directives for advance unit commitment and scheduling, and continuing compliance with dispatch directives. In addition, the ISO can designate plants that “must run” for reliability and pay them their incremental costs. However, these requirements differ immaterially from the previous procedures within utilities and in power pools. Even the proliferating regulatory interventions (e.g., “must-offer” obligations to bid all available capacity, and advance scheduling of deferrable maintenance) largely re-establish the comprehensive control by the utility or the power pool before restructuring – the notion that tight control is unnecessary was abandoned after the California crisis. Following the California crisis, FERC proposed its SMD that essentially replicates the design of the northeastern ISOs descended from previous power pools. From the viewpoint of a generator, SMD really has only one product, which is generation capacity that is mostly stable from day to day and in every hour, as are its other attributes such as location and ramp rate. The SMD requires that an IPP must provide each day for each generation unit a “three-part” bid that speciﬁes its start-up and no-load costs and its schedule of bids for energy. In addition, the ISO knows its location, maintenance schedule, heat rate, ramp rate, minimum and maximum generation levels, and other technical parameters, as well as its fuel constraints and commitments to bilateral contracts and to exports to other control areas. If one takes FERC’s SMD as the standard, then one can summarize a generator’s viewpoint rather simply. Operations are about the same as they would be in a vertically integrated utility or a power pool. Indeed, because many IPPs are subsidiaries of energy companies that own both generators and utilities, the organizational aspects are unchanged in some respects. But ﬁnancial matters are vastly different because remuneration derives entirely from market prices. The main effects are therefore strong incentives to minimize costs, and because they are completely exposed to the volatility of market prices, strong incentives to contract forward via bilateral contracts. This conclusion differs somewhat for other systems (e.g., NETA, NordPool, Germany, Australia) that rely less on optimization by the system operator and more on selfscheduling, and rely more on forward contracting than on spot markets. The main conclusion remains, however, that for generators it is the ﬁnancial implications of restructuring and liberalized markets that are most important. This accords with expectations that were raised when restructuring was initiated, since even then it was believed that operational aspects would not be changed materially if a system operator took over ongoing management of the grid, and possibly also spot markets. The results attest to the success in establishing well-functioning system operators (of various designs), and also to success in making proﬁt as measured by regional market prices the criterion of ﬁnancial performance.

EBL

1.4. The Unsolved Problems of Liberalized Markets This section compiles a summary of the problems that persist after the initial years of restructuring of the electricity industry, and the accompanying liberalization of wholesale and retail markets. The discussion is divided into two subsections. The ﬁrst summarizes the problems that cannot be solved efﬁciently by market processes. The second summarizes the problems that might in principle be solved by market processes, but that in fact current

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

49

designs have made little progress in solving adequately due to various practical aspects. In both cases these problems are deeply intertwined with the fundamental problem of how best to allocate risk bearing among market participants.

1.4.1. Problems not to be solved by markets This subsection focuses on electricity industry functions that require continuing regulatory interventions, regardless of market design. These functions including meeting transmission system requirements, maintaining a reliable grid, and guaranteeing universal service. 1.4.1.1. Meeting transmission system requirements The role of the transmission grid as necessary infrastructure is even more important in liberalized markets because it is the common highway for energy trading. An AC system is synchronized continuously over its entire span, and ﬂows must be adjusted continuously to control frequency, voltage, and line loadings. Flexible AC Transmission System (FACTS) technology may eventually allow ﬂows to be regulated by varying impedances, but most systems presently require operators to manage regional segments and to coordinate among them. The natural monopoly and economies of scale of transmission investments are therefore reinforced by the necessity of tightly coordinated operations. Recognizing this, many countries long ago established state-owned or regulated national transmission companies charged with ensuring adequate investments and ongoing operations. Those with privately owned or predominantly utility-owned transmission systems depend instead on regulation. Most simply provide for recovery of investment and maintenance costs, while the more advanced impose grid management charges and congestion charges and eliminate pancaking of transmission access charges. A fundamental innovation of restructuring, however, is to require open access and nondiscriminatory pricing according to the principles of common carriage – minimal requirements for liberalized wholesale markets. These measures, however, omit three fundamental requirements. One is efﬁcient management of grid operations. This is not assured by assigning the task to an ISO of the form that, in United States, is a non-proﬁt bureaucratic organization with diffuse incentives, unresponsive governance, and discretion restricted to explicit rules codiﬁed in its tariff. Performance-based regulation, as exempliﬁed in the United Kingdom’s regulation of the NGC, offers prospects for stronger incentives and greater discretion to manage grid operations ﬂexibly to cope with circumstances as they arise. The second fundamental requirement is planning of transmission expansion. Restructuring has weakened the integrated resource planning previously undertaken by utilities and in the United States left no agency responsible. This is especially serious now that liberalization has impaired coordination between transmission and generation investments. When generation investments are undertaken privately, it is imperative that transmission planning establishes reliable forecasts of the topology of the grid in future years so that decisions about generation investments can be adapted to the conﬁguration that will be in place during at least the ﬁrst 10 years of a plant’s life. The best plan will likely exploit the possibilities for substitution between generation and transmission that an integrated plan might achieve incompletely, but it is still necessary to establish reliable predictions of grid expansion. Regulatory authority is often required because private generation companies invariably prefer to build in a load pocket even if transmission expansion might be more efﬁcient.

EBL

50

Competitive Electricity Markets

The third fundamental requirement is ﬁnancial incentive for transmission investment. The regulated rate of return usually allowed for cost recovery of transmission investments is inherently vulnerable to the “Averch–Johnson effect” of favoring excessively capital-intensive investments (Sherman, 1985).8 This is especially true in liberalized markets because transmission projects must compete for funds with other projects, such as generation investments. A deeper problem, however, is that private incentives for transmission investments can differ substantially from social motives because the distributional effects can be substantial. For example, a retail utility in region A may want to fund a transmission line that enables it to purchase energy from region B where generation costs are lower. But if the effect of the new line is to equalize wholesale energy prices in regions A and B then an accounting of the aggregate social beneﬁt must also consider the higher energy prices charged to retail customers in region B, as well as the effects of price equalization on suppliers in both regions. There are also new motives for transmission projects undertaken to diminish the market power of local generators. 1.4.1.2. Maintaining reliable grid operations An important innovation of restructuring is the assignment of ongoing grid management to a system operator. The public good obtained from the transmission infrastructure depends continuously on protecting quality attributes (frequency, voltage, waveform), service reliability, and security against cascading failures. As mentioned previously, power markets are fundamentally incomplete due to disparities between technical constraints on supply (e.g., generator ramp rates, reactive power requirements) that affect reliability, and customers’ perceptions that they have no choice for power of the requisite quality. Rather than relying on markets, operating procedures rely on well-developed engineering principles to sustain quality and to protect reliability and security. A system operator obtains needed resources from spot markets (or “out of market” if necessary), and it supports and facilitates these markets, but ultimately it invokes command-and-control methods to ensure rigorous coordination. This reﬂects the reality that grid management assures physical feasibility, whereas markets determine mainly the ﬁnancial terms for settling transactions. There remains, however, considerable latitude for the system operator to affect the efﬁciency of congestion management and the efﬁciency of its spot markets. The performancebased incentives and wide discretion allowed NGC in the United Kingdom recognize this. Indeed, the reductions in NGC’s uplift and the increases in PJM’s congestion cost strongly support a conclusion that incentives for the system operator should be an important ingredient of regulatory policy. In contrast, in the United States the prevalent view that engineering “best practice” determines the ISO’s actions is the main justiﬁcation for ignoring incentives and relying on a bureaucratic non-proﬁt organization. FERC’s orders allow ITCs, like NGC, to operate within ISOs and RTOs so one expects that in the future their role may increase, and thereby incentives might be strengthened. However, a basic impediment to the ﬂexibility and discretion required for full efﬁciency remains the reliance in the United States on the rigid tariff prescribing the main aspects of each ISO’s operating procedures and market rules, as well as the prohibition against stakeholder involvement in governance.

EBL

8 Averch–Johnson effect (AJ effect), named after two economists who found that under rate-of-return regulation, if the allowed return is greater than the required return on capital, the ﬁrm will tend to over-invest in capacity beyond what is needed for economically efﬁcient production.

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

51

Among system operators, the designs of spot markets differ markedly. Their scopes range from a single balancing market to multiple forward and real-time markets. Their procedures range from simple market clearing to elaborate optimizations of all aspects simultaneously. And their ﬁnancial settlements range from energy payments to prices elaborately differentiated by time and location and differentiated among energy, various reserve categories, transmission, and in rare cases even payments for reactive power and other related products. There is no simple explanation for the huge difference between the United Kingdom’s NETA, which uses only bilateral trading and self-scheduling up to a few hours before the real-time balancing market, and the enormously complex markets of the ISOs in the United States. A few features are evidently relevant (e.g., the lower incidence of transmission congestion in the United Kingdom) but both evolved from power pools with somewhat similar procedures initially. A key difference is that, for complex reasons, the United Kingdom rejected the previous pool style of organization when it devised the simpler NETA, whereas at nearly the same time in the United States the experience of the California crisis motivated FERC to guard vigorously against a repetition by proposing its SMD. The result is that NETA relies on NGC to assure physical feasibility on a time frame of hours, while the ISOs establish physical feasibility day-ahead in consolidated markets for energy, reserves, and transmission (including unit commitments), maintain dispatch control throughout, and penalize real-time deviations. It remains unclear whether these and other system operators might converge to market designs with somewhat similar features. Continued experimentation with market designs might eventually produce convergence, but it is also possible that they will differ permanently due to initial conditions. Certainly in the United States the shock of the California crisis just 2 years after its ISO began operation has remained so vivid that comparisons with the performance of the NETA design are ignored. The crux of the difference is the considerable conﬁdence in the United Kingdom that NGC can manage physical feasibility and that bilateral trading sufﬁces for efﬁcient trading, while the accepted view in the United States is that physical feasibility must be established absolutely a full day ahead – for fear that any discrepancy will again open opportunities for gaming and abuses of market power – and therefore that the ISO’s markets must be multilateral, consolidated, and optimized. Differing experiences are also relevant to comparisons among other systems – NordPool, Germany, Australia, New Zealand – each of which has a market design whose unique features were heavily inﬂuenced by experience. However, an open issue for every market is the provision of adequate investment incentives to ensure long-term resource adequacy. A facet of these considerations is paramount when predicting the future role of system operations. One might make the case that markets in NordPool (established by the national transmission companies) and Germany (conducted by utilities) developed organically as energy trading grew, and in the United Kingdom the privately run markets for bilateral trading have developed vigorously after NETA began. But in other cases regulators and/or legislators have either chosen the market design or largely determined its main features, except for those aspects narrowly circumscribed by engineering requirements. Some may be good designs, and certainly others have been deﬁcient – notably the deep ﬂaws imposed on the California design by the legislature and the PUC. A basic issue to be resolved is whether in the future the spot markets of the system operator can and should cater to the commercial interests of the market participants the way other commodity markets do, or whether the peculiar technical requirements of power systems require that regulators have ﬁnal authority to specify the market design.

EBL

52

Competitive Electricity Markets

1.4.1.3. Guaranteeing universal service One part of retail service that is not entirely amenable to market processes is assurance of universal service. Extension of distribution lines to remote customers is now rarely a problem, so the main requirement is provision of service under standard terms and conditions. Regulators retain authority to deﬁne standard service plans, and to ensure that there is a POLR. The regulatory compact was the means in the era of vertically integrated utilities, but other means are now possible. The POLR was selected via a procurement auction in some cases where the utility’s incumbency advantage was not so great as to exclude effective competition from other LSEs. But competitive procurement with ﬁxed remuneration runs the risk that a POLR that is an unregulated LSE becomes insolvent when wholesale prices rise. As a result, the dominant mode is to rely again on the regulatory compact, so that utilities provide basic service and recover their costs from retail rates that are nearly level over time. This mode is deeply at odds with the initial view of what liberalization of retail markets could accomplish. Ideally, retail liberalization implies that retail customers pay the hourly wholesale price for energy (in addition to other charges for distribution) but they also purchase ﬁnancial hedges against price volatility to the extent they prefer. This ideal cannot be realized anytime soon because metering is still primitive and the markets for ﬁnancial hedges are undeveloped, but more basically this ideal is impractical for the (mostly small residential) customers most affected by universal service, even if ﬁnancial hedges were subsidized. Relying instead on utilities, and using the regulatory compact to assure cost recovery over time, is likely to remain the only practical solution for decades. This poses two basic problems. The ﬁrst is that the “core” of customers opting for basic service is inherently unstable. Those who opt out when wholesale prices are low are equally motivated to opt again for basic service when wholesale prices are high. Charges might be imposed on those who leave the core and again on those who return, but exactly how such a system would ensure ﬁnancial stability through prolonged swings of wholesale prices is not well understood and has not been fully developed. The second basic problem occurs even if those who opt out of the core never return. The customers who are least costly to serve (e.g., those with ﬂat load proﬁles) are precisely those who will be offered the most attractive terms by IPPs and non-utility LSEs, and therefore they are the ones most likely to opt out of the core. This leaves in the core only the customers most expensive to serve, implying that their retail rates (even if leveled over time) will rise as the core is depleted of the more proﬁtable customers. This scenario is entirely realistic since it merely repeats the dire experience of those utilities that, late in the regulated era, were subject to bypass by proﬁtable industrial customers who opted for self-generation or co-generation, or later, direct contracting with IPPs.

EBL

1.4.2. Problems not yet solved by markets The discussion now turns to a different class of problems, those that liberalized markets were initially believed to solve, but in fact have not. These include assuring adequate generation resources and beneﬁting retail customers, an issue also addressed in Chapters 9–13 of this volume. 1.4.2.1. Assuring adequate generation resources The generation sector was thought to be the most amenable to market solutions. There is some evidence that operating efﬁciency has improved at divested plants, as measured by labor and fuel (heat rate) inputs. Other aspects of operations are less clear in those systems

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

53

that exclude self-scheduling for those plants not committed to bilateral contracts. Those ISOs in the United States that use variants of FERC’s SMD are most extreme in abandoning market-clearing processes in favor of comprehensive optimization, including unit commitments and scheduling of energy generation and reserve assignments. Although these ISOs settle accounts using prices computed as the marginal costs of unbundled products, their operations consist essentially of centralized allocations of available generation and transmission capacities. These procedures differ little from those used previously in vertically integrated utilities and power pools. These ISOs in the United States might eventually resemble the more decentralized systems in other countries if, as FERC intends, bilateral contracting becomes more pervasive. At present, however, the fraction of plants scheduled centrally is substantial and stable, and in many ways FERC’s insistence that each ISO must ensure day-ahead physical feasibility necessitates centralized unit commitment and scheduling, and continuing dispatch control until real-time. The mix of generation investments has been problematic in most ISOs. An encouraging sign is that many systems now have substantial dispatchable loads that compete with generators in reserve markets, and increasingly the demands of industrial customers respond to real-time prices. The number of interruptible or curtailable retail service contracts remains comparable to the pre-restructuring period, but in general the utilities’ active promotion of demand-side management declined after restructuring as regulators withdrew the subsidies and incentives previously provided. Regulators have continued some subsidies of generation from renewable sources, often by using market mechanisms such as auctions, but also by innovative schemes such as the tradable “green certiﬁcates” introduced in Europe. In addition, renewable sources increasingly compete on equal terms, albeit with inherent disadvantages because generation need not be dispatchable and can be intermittent (e.g., wind and solar). Now the mix of generation investments is also affected by requirements for tradable permits for emission of pollutants, such as sulfur dioxide and nitrous oxides. The ratiﬁcation of the Kyoto Treaty will in future years lead to comparable requirements for emissions of carbon dioxide and other greenhouse gases. The most severe problem concerns investments in peakers (e.g., CTs) that routinely provide fast-response ofﬂine reserves and occasionally generate to meet peak loads when prices exceed their high marginal costs. In other countries such as Australia, bid or price caps are sufﬁciently high that peakers obtain substantial revenues in times of supply scarcity (and encourage demanders to obtain ﬁnancial hedges). While a high price cap is arguably the best solution, the low caps in the United States curtail CTs’ revenues from occasional generation and force them to rely mainly on payments for reserve assignments. However, centralized unit commitment by the ISO, usually undertaken to strengthen reliability, has the side effect that it depresses the price for ofﬂine reserve capacity, because the minimum operating level of a newly committed unit tends to create an excess supply of spinning reserve. The usual motive for extra unit commitments is under-scheduling of loads in the dayahead market, compared to the ISO’s forecast of the next day’s peak load. This problem is acute in New England because its ISO needs substantial ofﬂine reserve capacity to meet its unusually large contingencies, but reserve prices are insufﬁcient to cover the carrying cost of a CT. The ISO there is exceptional for addressing this problem by allowing CTs to capture scarcity rents. One device is a separate semi-annual market for ofﬂine reserves, and a second is a contingent capacity payment awarded to all available capacity whenever reserves are in short supply in real-time operations.

EBL

54

Competitive Electricity Markets

These measures are indicative of a general trend toward special provisions intended to obtain sufﬁcient supplies of reserves in daily operations and more elaborate efforts to assure adequate generation resources. The plain fact seems to be that adequate reserve capacity – both fast-response reserves in the short term and total available capacity in the long-term – is not assuredly provided by the ﬁnancial incentives derived from wholesale markets. Reserve capacity can be insufﬁcient when prices in these markets are depressed by regulatory interventions, as in the United States where low price caps and residual unit commitments for reliability affect revenues obtained by CTs. But the basic economic problem is that individual customers have insufﬁcient incentives to pay for reserve capacity (and its chief attribute, a high ramp rate) that is idle most of the time and that serves mainly to provide the public goods of system reliability and grid security. Customers and regulators are acutely sensitive to these matters when prices are high because loads are high or supply is scarce, but ordinarily their ﬁnancial incentives are myopic compared to the integrated resource planning previously conducted by utilities. A chief motive for separating the organizational components of vertically integrated utilities was to establish stronger incentives for efﬁcient investments in generation. Rather than assured cost recovery obtained by a utility, a private investor in generation obtains the subsequent rewards and bears the risks that follow from its decision. But the evidence is mixed as regards investment. Success stories in countries like Australia and in ERCOT, which is not subject to FERC jurisdiction, are matched by a bleaker picture in other parts of the United States and in Europe. The insufﬁciency of new investments was initially attributed to prolonged regulatory uncertainty, compounded by the demise of new contracts for QFs, including both co-generation and those using renewable energy sources. Then high wholesale prices in the period surrounding the California crisis stimulated substantial investments by IPPs, mainly in combined-cycle gas-ﬁred plants. But this was followed by rather low prices that resulted in ﬁnancial distress for many IPPs and bankruptcy for some. This could be a sign that boom and bust cycles are endemic in the generation sector, as they are in some other commodity industries. But if so then it reﬂects a basic deﬁciency in implementing liberalized markets, which we now address. Restructuring was based on two premises implied in the seminal book by Joskow and Schmalensee (1983). One was that the vertical integration of utilities could be replaced by bilateral contracts between generators and large customers, or with retail utilities and other LSEs, assisted by multilateral markets for spot trading. The other was that generators could obtain capital on comparable terms directly from ﬁnancial markets without relying on inclusion within a regulated utility with its assured cost recovery. Even California, which prohibited long-term bilateral contracting by utilities during the 4 years allowed for recovery of stranded costs, presumed that contracting would ultimately prevail. Contracting was certainly the dominant mode initially in those systems that used vesting contracts to smooth the ﬁrst years of transition. In some systems (e.g., NETA, NordPool, Australia) both long- and mid-term contracting are vigorous, and others (e.g. France, Alberta) stimulated contracting by auctioning power procurement agreements (PPAs) sold by incumbent owners of large amounts of generation who were not required to divest ownership and management of their plants. But experience has shown that there are basic problems involved in extending the role of contracting. As described earlier, sellers and buyers have a natural incentive to insure others against volatile spot prices, since these are just ﬁnancial transfers between them. The prevalence of mid-term CFDs reﬂects this mutual incentive for price insurance. Other stipulations involve the contract duration and the quantities involved. Utilities and other LSEs are hesitant to contract long-term for ﬁxed quantities when their customers can switch

EBL

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

55

to alternative providers, while generators prefer such contracts. Similarly, utilities and other LSEs want to adjust contracted quantities to contingencies affecting loads whereas generators do not. These considerations may account for the smaller role of contracting in the United States. The creditworthiness of the contracting parties also seems to have been an impediment in the United States, both in the California crisis when the utilities were ﬁnancially distressed and later when there were several bankruptcies of major generation companies. Utilities’ regulators are wary of default on supply contracts, and they are also reluctant to approve utilities’ purchases of purely ﬁnancial short-term contracts (e.g., futures contracts, as opposed to forward contracts that entail delivery). Indeed, most PUCs prohibit their inclusion in cost recovery on the grounds that they amount to speculation or arbitrage, a view that is evident also in federal legislation that requires separate and more intrusive regulation of exchanges for commodity futures. Also relevant is the slow development of innovative forms of forward contracts. For instance, an option contract can enable an LSE to “call” for the delivery of power supply at a pre-speciﬁed strike price in contingencies when the spot price exceeds the strike price. An option contract enables a buyer to hedge against high prices without exposure to the “volumetric” risk that the contracted quantity exceeds the amount required to serve its load. A multitude of other contract forms are also possible; e.g., a “spark-spread” contract enables a seller to hedge differences between fuel and power prices. Predictions that contracting would develop vigorously, and fund investments in generation, were predicated on the assumption that adequate contract forms could be developed to overcome the concerns of sellers and buyers, and that prudency reviews and creditworthiness would not be major problems. Predictions that contracting would replace vertical integration have been only partly realized because in fact the impediments to contracting interacted with the other major premise of restructuring – namely, that generation companies could obtain capital directly from loans and from ﬁnancial markets for bonds and equity shares, much like suppliers in other deregulated industries (transport, telecommunications, gas). In the United States the chief models were gas transmission companies, which typically obtained funds at low cost because before construction of a pipeline most of its capacity was pre-sold as ﬁrm transmission rights under long-term contracts, often for durations as long as 20 years. Some IPPs have pre-sold capacity for durations long enough to provide security for loans to fund investments, but the overall pattern is mixed. In California, over 90% of the 8 GW of new capacity (an increase of 20%) installed in the years after the crisis was being ﬁnanced by the long-term ﬁxed-price, ﬁxed-quantity contracts that the state purchased when it intervened. The state later found itself burdened with contracted prices and quantities that turned out to be excessive. This salutary lesson now encourages PUCs to restrict the durations of utilities’ contracts for resource adequacy, and further, to impose stringent conditions for taking over the plants of a supplier in default. Default risks are substantial because of the inherent volatility of wholesale prices, even over extended periods. Recent years of sustained low prices have jeopardized the ﬁnancial viability of even the major generation companies, especially those active in energy trading with exposed positions (e.g., Calpine, Dynergy, Enron, Mirant, Williams). In principle, a generation company could be fully hedged by long-term ﬁxed-price, ﬁxed-quantity contracts and therefore be immune to long swings of wholesale prices, and if its counterparty is a utility with assured cost recovery, then its credit is essentially comparable to the utility’s credit. But complete hedging is rarely available, and the contract duration is typically no more than 3 years, a minor fraction of the life of the plant. Financial

EBL

56

Competitive Electricity Markets

hedges (e.g., futures contracts) are intrinsically short-term because no counterparty to a long-term ﬁnancial contract is creditworthy unless it has physical resources as a collateral. A forward contract for physical delivery over a long duration can be viable for the seller, but no LSE other than a utility is a creditworthy buyer. As mentioned earlier, the LSEs, utilities, and industrial customers must guard against quantity risks. An option contract mitigates the buyer’s quantity risk, but equally it imposes on the seller the quantity risk that over long periods with low prices the option will not be called. Because the option will be called only when spot market prices are higher than the strike price, the seller serves mainly as insurer of the buyer with no comparable insurance for itself. The result of these difﬁculties is that two basic premises of restructuring have been fulﬁlled only partly. Contracting is not as pervasive as expected, and contracting obtains only part of the ﬁnancial advantages of vertical integration. Most important is that IPPs are risky companies and therefore they have higher costs of capital – and during the recent period of sustained low prices, some of the most prominent are ﬁnancially distressed, bankrupt, or reorganized after bankruptcy. The ultimate consequence is that investments in new generation by IPPs are substantially impaired. For example, Calpine obtained regulatory approvals for siting and construction of three new plants in California for which it did not obtain investment funds. 1.4.2.2. Beneﬁting retail customers Restructuring and liberalization of retail markets were expected to beneﬁt customers substantially. For example, California’s announcement of its decision extolled at length the merits of differentiated services that customers would obtain at competitive prices from LSEs and from utilities freed from the conﬁnes of standard service plans. Much of the gain was supposed to come from service plans that rewarded customers’ efforts at demand-side management of their energy consumption, complemented by integration (called “convergence”) of energy services with other basic services – especially telecommunication and (remote or local) automated control of appliances, which have yet to materialize to any great extent.9 Large industrial and commercial customers certainly obtained advantages from new freedom to contract bilaterally with IPPs and to purchase directly from retail energy providers (REPs) or wholesale markets.10 But LSEs made slight inroads into commercial and residential retail markets in the United States, and subsequently, those who survived ﬁnancial stresses during and after the California crisis withdrew from these markets.11 The utilities retained 90% or more of the residential market in most states simply because few customers considered switching energy suppliers. An exception is ERCOT where retail competition is thriving and 50% of the load has switched away from their local utility suppliers. A signiﬁcant fraction of commercial customers chose service from LSE’s whose rates were gauged to each customer’s load proﬁle, but this was perhaps the only signiﬁcant differentiation and succeeded mainly because it provided lower prices to customers less

EBL

9

California’s 1994 restructuring decision devoted two pages to a section entitled “The Convergence of Telecommunications with Electric Service Promotes Direct Access” (pp. 21–23). 10 For example the University of California at Berkeley signed a supply contract with ENRON that guaranteed a minimum of 5% saving over the PG&E regulated rate. During the energy crisis ENRON attempted to renege on that contract and force the university to return to PG&E, but lost its bid in court. 11 California suspended “direct access” at the end of the crisis, and the suspension continues still.

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

57

costly to serve. More elaborate differentiation was impeded by the prevalent absence of interval meters, and the costs of more elaborate meters and billing procedures. Prior beliefs that customers would actively monitor and control usage to reduce their costs under new service plans were falsiﬁed by the same practical obstacles that had limited the results obtained previously from utilities’ demand-side management programs. Equally falsiﬁed was the expectation that a majority of customers would willingly bear short-term price volatility (or could and would purchase ﬁnancial hedges) in order to obtain lower prices on average. In fact, when wholesale prices were passed through to customers (in San Diego in 2000 as the California crisis began), the howls of outrage from customers prompted immediate intervention by regulators who canceled provisions allowing the utility to pass through its wholesale procurement costs.12 A particular example of misguided expectations was the prediction, obtained from surveys of mainly residential customers, that 30% would willingly pay a premium of 10% for “green” power from renewable sources. In fact, those subscribing to green power never exceeded 3%. More successful liberalizations of retail markets do not rely on visions of elaborately differentiated services. Instead, they concentrate on competitive retail markets as means of lowering prices for standard service plans, and for enabling negotiated rates based on a customer’s load proﬁle. Alberta’s auction of PPAs, for instance, enables LSEs who purchase them to compete against the incumbent utility. Texas encouraged the entry of competitive REPs by requiring utilities to refrain from price competition with the REPs at prices below a “price to beat” until 2007 or until the utilities’ market share dropped to 40%, whichever occurred ﬁrst. A basic lesson from recent years of liberalized retail markets is that only the market for large industrial and commercial customers is sufﬁciently developed presently to beneﬁt fully from liberalization (see Joskow, 2005). Extending retail competition to other market segments requires major advances in infrastructure (especially metering), and fundamental redesign of differentiated services. Also required are measures that allow LSEs to enter successfully, to establish signiﬁcant market shares, and to remain ﬁnancially viable through periods of high prices. The reason is, in part, that they insure their customers against wholesale price volatility, although on the timeframe of a month or a year, which is shorter than utilities provide. Most fundamental is the lesson that most customers in the commercial and residential segments are deeply averse to price volatility, and equally reluctant to undertake the measures required to continually monitor and control usage. These customers will remain in the “core” served by the utility using fairly standard service plans (only slightly differentiated, say, by peak and offpeak periods) with rates leveled over extended periods. Only if there are major advances in providing them with ﬁnancial hedges against price volatility, and with compensation for service interruption or curtailment, are these customers likely to depart from the core.

EBL

12

Similarly, Ontario canceled liberalization of retail markets after prices rose 30% in the ﬁrst months. Customers’ attitudes about pass-through of wholesale electricity prices have no easy explanation. During the California crisis there was widespread anxiety about their monthly bills, but in fact throughout the crisis electricity rates were regulated and rigidly ﬁxed. The anxiety can be explained in part by the fact that gas prices were passed through to customers and bills from PG&E and SDG&E included charges for both electricity and gas. But why many customers did not notice that it was higher gas prices that accounted for their higher monthly bills is mysterious. The new role of electricity as necessary for commerce and for personal well being – even TV – may be as good an explanation as any for these attitudes.

58

Competitive Electricity Markets

1.5. The Allocation of Risk Bearing in Liberalized Markets This section analyzes risk management in the electricity industry and indicates the continuing role of regulatory policies and regulated retail utilities in mitigating risks for a segment of retail customers. Rather than the simplistic version of liberalization that aims simply to break apart the vertically integrated utilities, and to unbundle the products traded in markets, there is a continuing role for utilities in providing inter-temporal smoothing of retail rates, and in lowering the cost of capital by reducing their ﬁnancial exposure.13 This role requires redesign of utility investments and contracting within the context of liberalized wholesale markets, and redesign of retail rates within competitive retail markets. First reviewed are the risks peculiar to the industry, and then the institutional arrangements that have been or can be used to obtain an efﬁcient allocation of risk bearing. 1.5.1. The basic risks affecting the electricity industry Much of the theory and practice of risk management is based on diversiﬁcation. Insurance companies are ﬁnancially viable because they aggregate many small independent risks to life, health, property, etc. They can operate with relatively small reserves of equity capital because their aggregate risk is small on a per capita basis, that is, when divided relatively equally among all participants. Another form of diversiﬁcation is seen in a mutual fund. Its earnings are less volatile than the earnings of the companies whose shares it owns because its portfolio is divided among the shares of many companies. Again, the aggregate is less risky per dollar invested because the companies’ earnings are imperfectly correlated. Diversiﬁcation is also very important in electricity markets. For instance, the aggregate load on a per customer basis is more stable than the loads of the customers individually because the aggregate is composed of individual loads that are imperfectly correlated. Thus, for each hour of the next day, a system operator’s day-ahead prediction of the average load per customer is more accurate than its predictions of the loads of individual customers. Similar considerations apply to other contexts; e.g., the system operator uses the transmission system to compensate for equipment failures in one area by drawing on resources in other areas. Analogously, the system operator can meet a high load in one area by drawing on resources in other areas that are not affected by the same weather conditions. Energy trades also take advantage of differences between seasonal variations among regions; e.g., California buys power from the Northwest to meet summer daytime air conditioning loads, and sells power to the Northwest to meet winter nighttime heating loads. The mix of generation technologies that is most cost-effective in meeting a particular load-duration proﬁle is really the solution to a risk management problem in which diversiﬁcation of generation investments is optimized. Both spatial and temporal diversiﬁcation, however, are inherently limited in the electricity industry. Investments in generation and transmission facilities require years to complete. After construction, they are irreversible, specialized, immobile, and long lived.

EBL

13

One problem underlying the spotty record of power industry restructuring so far is that the economic theory used to justify functional unbundling of utilities has not proved as useful as originally expected. This theory, which emphasized the importance of transaction costs, depended largely on assumptions that have become outdated because of innovations in technology and market design. It now appears that many of the economic beneﬁts sought through unbundling can better be attained through wider use of risk management contracts, obviating a compelling reason why restructuring should begin with the irreversible task of vertical unbundling of the supply chain.

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

59

In contrast, loads vary greatly on much shorter time frames. Some variation in loads is predictable and cyclical, such as the typical variations over the hours of a day, and over the seasons of the year, so the mix of generation technologies can be designed to minimize total costs over the cycle. On the supply side too there are regular patterns of downtime for maintenance, and to a substantial extent the average frequency of equipment outages is predictable. But the electricity industry is also affected by large and relatively unpredictable variations that occur over wide regions and/or over long timescales. These bring risks that cannot be mitigated by various strategies of diversiﬁcation based on averaging over people, locations, or time in a cycle. These are called non-diversiﬁable or “systemic” risks. Among systemic risks, the most extreme event is collapse of the grid due to cascading failures. Systems that rely heavily on hydroelectric sources are vulnerable to prolonged droughts that curtail water storage behind dams, and those that rely on fossil fuels are vulnerable to eras of high prices. Over long periods, load patterns trend away from the aggregate load-duration proﬁle and the spatial distribution used initially to justify generation and transmission investments. Technical change can also render a generation plant or transmission line inefﬁcient compared to subsequent investments in newer technologies. For a casualty insurer, the analog of a systemic risk is the rare storm (e.g., hurricane) or an earthquake that devastates an area and requires that compensation be paid simultaneously to many victims of the same event. In other words, the injuries to insured customers are perfectly correlated due to their common dependence on the single event of the storm. Because a single event can exhaust its ﬁnancial reserves, a casualty insurer often excludes coverage of systemic risks. For instance, a farmer can purchase insurance against damage to crops by hail (which occurs locally and brieﬂy) but cannot purchase insurance against drought (which is widespread and prolonged). Casualty insurers typically diversify further by re-insuring a portion of their risks with other companies that specialize in aggregating calamitous risks on a global scale. Such strategies might conceivably be invoked in the electricity industry, but presently the scale of the ﬁnancial risks of major events in the electricity industry is so large, and can extend over such long times that it is often the state that ultimately bears the cost – as when the state of California intervened to purchase power for utilities during the crisis. Systemic risks are addressed in many different ways. Provision of reserve generation capacity is the ﬁrst defense: it is costly to build and maintain generators that are idle most of the time, but over the long run their ultimate value is realized in the occasional events when they are called to produce power. As with an insurer, a second defense is a reserve of equity capital that can be drawn down to pay extraordinary expenses. Again, it is costly to retain ﬁnancial reserves that, in effect, are idle until used to meet unexpected expenses. Like an insurer, a utility can use the third defense of re-insuring its ﬁnancial risks in various ways, such as long-term contracts that transfer some portion of its risk of high wholesale prices to its suppliers, and retail service plans that transfer some risk to customers. Risk associated with moderate weather ﬂuctuations can also be diversiﬁed through instruments such as weather derivatives that enable risk sharing among industries that are affected by weather in complementary ways. But some entities must ultimately bear the costs of rare extreme events, and in the electricity industry the prospects that ﬁnancial contracting can entirely diversify systemic risks are limited. The problem stems partly from the attributes of physical assets, which are built slowly, and are irreversible, specialized, immobile, and long lived, and, thus, largely inﬂexible in dealing with contingencies that occur on large scales of space and time. A contract that transfers risk from a regulated utility to an unregulated generation company leaves the company exposed to the risk but with limited ﬂexibility, since it cannot redeploy or

EBL

60

Competitive Electricity Markets

quickly expand its assets to cope with events as they develop. Ideally, a long-term ﬁxedprice contract protects the utility against sustained high spot prices (while forgoing high proﬁts for the company), and protects the company against sustained low prices (while foregoing lower procurement costs for the utility). The advantages for the company are inherently greater since its investments require irreversible commitments measured in decades, while the utility’s advantages are conﬁned to the durations of extreme events. But this mutual insurance against price variations is not sustainable over prolonged periods that jeopardize the ﬁnancial viability of either party – as evidenced by the bankruptcies of generators in the United Kingdom and the United States during the recent period of low prices, and before that the ﬁnancial distress of utilities and other LSEs during a period of high prices. The problem also stems partly from high correlation between prices and quantities; i.e., prices are high when loads are high. A contract that provides price insurance is mainly ﬁnancial, since it ﬁxes the terms of trade, and brings mutual advantages to buyer and seller by eliminating price volatility. But a contract ﬁxes the quantity to the beneﬁt of the seller only by removing the buyer’s ﬂexibility in procuring the amounts needed to meet its load in each event. If the amount is too large or small, then the surplus or deﬁcit must be corrected by spot sales or purchases. Basically, the seller wants to insure its ﬂow of net revenues (especially if it is burdened by debt) and the utility wants to insure its total cost of procuring supplies to meet its varying load. Some contracts address these considerations directly, notably tolling contracts (sometimes called “virtual capacity” contracts) and some PPAs, in which essentially the seller is remunerated continuously for its capacity availability and operation. Meanwhile, the buyer dispatches the plant as needed and pays variable generation costs, chieﬂy for fuel. This achieves the primary goal of restructuring, which is to make generation companies bear the consequences of their investment decisions and to strengthen their incentives for efﬁcient operating practices, while also enabling the utility to obtain scheduling ﬂexibility. Restructuring assumed that the strategies described above – physical reserves, ﬁnancial reserves, and contracting – would sufﬁce in the new era of liberalized markets. Each strategy has limitations and costs that have became clearer as experience accumulated. Remarkably, restructuring has often abandoned the traditional means of risk management, namely cost-of-service regulation of utilities. Risk was borne ultimately by retail customers, but only by amortizing recovery of a utility’s accumulated costs over time so that its retail rates were substantially level. This strategy uses diversiﬁcation in two ways: it distributes risk bearing widely among customers and over time. It also removes most risk bearing by the utility and its suppliers so that their capital costs are low, and eliminates the creditworthiness problem. It depends implicitly on the good faith and credit of the state in fulﬁlling the regulatory compact; indeed, it may be that the state is the only entity that can provide credible assurance that costs will be recovered later as promised. This strategy also has limitations and costs. These became very clear at the time of restructuring, since regulators had often seen evidence that utilities’ incentives for cost minimization were weak, and investments were rewarded on the basis of “tonnage of money invested.”14 California’s initial consideration of how to restructure proposed a scheme in which utilities would continue to serve core customers much as they had

EBL

14 This quote is from the February 1993 staff report to the California Public Utilities Commission by its Division of Strategic Planning, “California’s Electric Services Industry: Perspectives on the Past, Strategies for the Future”, page 100.

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

61

previously, while allowing non-core customers to purchase their power supplies directly.15 Although California rejected this scheme, other states retained a central role for utilities and continued the practice of amortizing costs over time to level retail rates. This necessarily implies that core customers ultimately pay the full cost of the services obtained, and in particular they collectively bear the systemic risk that the cost of service for the core will be high due to extreme events and long-term trends. 1.5.2. Institutions for risk bearing in the electricity industry Restructuring was a response to many considerations. One was the reduced role of economies of scale in generation as smaller combined cycle units became cost-effective. This and other technical advances obviated the dominant role of retail utilities on the supply side and opened prospects of bilateral contracting between IPPs and large industrial and commercial customers. For those systems not already organized as power pools, there were potential operational gains from regional operations and trading based on open access to transmission. Naïve expectations that service differentiation would proliferate amid vigorous retail competition were realized only partly, and mainly for large industrial and commercial customers. Some PUCs emphasized reduced costs of contentious procedures and litigation required to implement cost-of-service regulation, but in fact this outcome was precluded by the continuing dominant role of retail utilities in serving core customers. A primary goal of restructuring was to strengthen incentives for efﬁcient operational and investment decisions. Cost-of-service regulation is inherently a kind of insurance for utilities, since it guarantees to a utility that its costs accepted as prudent and accepted into its rate base are eventually recovered in full from retail rates on an amortized basis that includes the cost of capital. Insuring utilities’ cost recovery was very effective in reducing the cost of capital, since their bonds and shares carried negligible risks of default and provided steady payments of interest and dividends. But inevitably, insurance dilutes incentives, since a utility does not bear the costs that result from its investment decisions and operating practices. Restructuring apparently succeeded as regards cost reduction in daily operations, but it has had mixed success in investments, and the unfavorable consequences for the riskiness of utility shares was not anticipated.16 Restructuring envisioned that supply-side contracts with independent suppliers, supplemented by spot market purchases, would supplant in-house generation by vertically integrated utilities. And on the demand side, differentiated service plans would include risk bearing by customers (possibly hedged by ﬁnancial instruments) on a more nearly current basis than the previous regime of rates leveled over extended periods. Also expected was that customers’ loads would become more sensitive to market prices. These two developments on the supply side and the demand side were

EBL

15

The reasons for rejecting this alternative are not entirely clear. One explanation is that California rates had already reached averages 30% to 50% above the national average while relying on costbased regulation, and they might rise further if the most proﬁtable customers opted out of the core. But the PUC’s decision focused also on predictions that complete liberalization would bring beneﬁts from service differentiation among competing LSEs that in fact did not materialize. In 2004 the PUC reinstated the policy of promoting utilities as the provider of services for core customers. This reversed the decision made a decade earlier. 16 Regulators did not expect utilities to be in ﬁnancial jeopardy, but equity markets reacted differently. After the California PUC’s decision in 1994 prices of the utilities’ shares declined about 25% over the next few months.

62

Competitive Electricity Markets

expected to leave retail utilities with moderate commercial and ﬁnancial risks, much like other ﬁrms in commodity industries – analogies to successful deregulation of the transport, telecommunications, and gas industries were often cited. The limits to the success of restructuring regarding investments include effects of imperfect planning, especially as regards the mix of technologies and the provision of adequate reserves. But the main limit is due to imperfect contracting that leaves generation companies exposed to substantial risk and therefore required to pay higher costs for capital. The failure of demand-side innovations to develop is now viewed as fundamental. This perception may change gradually through a long process of developing retail markets – expansion of sophisticated metering and redesign of marketing strategies – but the basic impediment is the absence of adequate ﬁnancial instruments for small customers to hedge against price volatility. A utility retains its obligation for universal service as the provider-of-last-resort, and its standard service plans include leveled rates that effectively insure against short-term price volatility. Most small customers therefore choose to remain in the core served by the utility. (In the United States, those who opted for service from alternative LSEs were sent back to their utilities for default service when wholesale prices rose in the period 2000–02.) There is no evident substitute now for the retail ﬁnancial services of utilities, and there may indeed be no credible substitute for the state’s guarantees of universal service and of cost recovery via leveled rates. In retrospect, the anticipated gains from service differentiation for small customers must presently be seen as secondary compared to the gains from leveled rates for those in the core. From an economic viewpoint, there are persuasive arguments that the most efﬁcient allocation of risk bearing in the electricity industry has core customers paying leveled rates for amortized cost recovery. Except for some economically disadvantaged customers, they should pay the full cost of service in the long run because costs vary with usage. Since electricity is used universally, they should pay directly for service rather than rely on distortionary taxation by the state to cover deﬁcits. Some industrial and commercial customers can bear short-term price volatility without difﬁculty, and therefore they can pay spot prices and/or contract directly with suppliers. But if other customers are deeply averse to short-run volatility then ideally they should pay level rates that recover their costs over time. Inter-temporal smoothing of rates might be achieved by well-developed markets for ﬁnancial instruments for hedging against price variations and for insurance against the consequences of curtailed service. Alternatively, cost-of-service regulation provides smoothing of rates. The choice between these two approaches depends on how seriously systemic risk limits the market for ﬁnancial instruments, and how serious are the deﬁciencies of cost-of-service regulation in providing strong incentives. The defects of cost-of-service regulation were well known before restructuring, but the slow and ultimately inadequate development of competitive markets for ﬁnancial instruments was not expected. It was also thought to be a secondary consideration, since customers retained the option to rely on core service, and in any case little or no attention was given to the problem of systemic risk. For instance, documents and orders of the California PUC and state legislation mention only short-term price variation, with no recognition that extreme events like those that initiated the later crisis might affect the restructured industry. Most ﬁrms in the ﬁnancial services industry were well aware of the threat posed by systemic risks. They hesitated about offering long-term ﬁnancial instruments because they had no physical hedges. A ﬁrm that offers ﬁnancial hedges against high wholesale prices runs the risk of ruin unless it can compensate by simultaneously proﬁting from selling power at high prices. Several ﬁrms within the energy industry became active traders and

EBL

Reevaluation of Vertical Integration and Unbundling in Restructured Electricity Markets

63

arbitrageurs (including Dynergy, Enron, Mirant, and Williams) and they claimed that their portfolios of long and short positions were well hedged against extraordinary events. But in fact, among the major trading companies in the United States, the only ones solvent after the California crisis and ensuing events at the national level were those that had the foresight to liquidate their positions and close their trading operations in the early months of the calamity. A summary view of these deﬁciencies on the demand and supply sides is that both are instances of insufﬁcient development of auxiliary markets for ﬁnancial instruments and contracts that hedge against risks. Diversiﬁable risks are allocated inefﬁciently when ﬁnancial markets are poorly developed, and more seriously, non-diversiﬁable risks can jeopardize the entire industry as ﬁnancial distress affects many participants. Dire scenarios were not envisioned when restructuring began, but they became worrisome concerns after episodes in several countries, and then became crystal clear during the California crisis, and later in some other countries such as New Zealand. The view now is that even after restructuring and liberalization there remains a valuable role for retail utilities that use ﬁnancial reserves from capital markets to smooth cost recovery over time for those customers who opt to remain in the core. The importance of strengthening incentives is also recognized, and therefore new regulatory policies are required. Remuneration of utilities via simple cost-of-service regulation must be replaced by a scheme that enables a utility to insure core customers against short-term price volatility, while also rewarding the utility for efﬁcient operations. Chao et al. (2006) address these matters in detail, including the role of performance-based regulation of utilities.17 1.6. Conclusions

EBL

The argument for vertical integration in the electricity industry and also the argument for restructuring based on unbundling of its products and organizations in favor of market mechanisms are both deﬁcient. The notion that all is needed is unbundling of the electricity supply chain and establishment of efﬁcient short-term trading institutions, while longterm contracting and markets for ﬁnancial risk management instruments will emerge spontaneously, was naïve. In retrospect, cost-of-service regulation and vertical integration of generation and retail service continues to be a powerful means of risk diversiﬁcation. The extremes of vertical integration and liberalized markets are inferior to a balanced mixture of the two approaches. While unbundling may beneﬁt large industrial and commercial customers that are able to absorb the inherent risks in the electricity supply chain, efﬁcient management of these risks requires that restructuring retains universal service for the core of non-industrial customers who rely on regulated rates smoothed over time to recover the costs of service. Acknowledgment The chapter is based on research sponsored by Electric Power Research Institute. The opinions expressed in this chapter are those of the authors and do not represent the positions of any of the organizations with whom the authors are afﬁliated. Any errors and opinions are solely the responsibility of the authors. 17 In its 1994 decision the California PUC explicitly provided for performance-based regulation of utilities that continued to serve core customers, but this was not implemented due to the restrictions imposed to enable utilities to recover the “stranded” costs of previous investments.

64

Competitive Electricity Markets

References Barker, J. Jr., Tenenbaum, B., and Wolf, F. (1997). Governance and regulation of power pools and system operators: An international comparison. World Bank Technical Paper No. 382. Blumstein, C., Friedman, L., and Green R. (2002). The history of electricity restructuring in California. CSEM Working Paper 103, Berkeley, CA: University of California Energy Institute. Bonbright, J.C. (1961). Principles of public utility rates. Public Utilities Reports, Incorporated. Borenstein, S. (2004). The long-run effects of real-time electricity pricing. CSEM WP-133, Berkeley, CA: University of California. Borenstein, S., Bushnell, J., and Wolak, F. (2002). Measuring market inefﬁciencies in California’s restructured wholesale electricity market. CSEM Working Paper 102. Berkeley, CA: University of California Energy Institute. California Public Utilities Commission (1993). California’s electric services industry: perspectives on the past, strategies for the future. San Francisco, CA. Chandler, A.D. (1969). Strategy and Structure: Chapters in the History of the Industrial Enterprise. The MIT Press. Chao, H-P, Oren, S.S., and Wilson, R.B. (2006). Alternative pathways to electricity market reform: risk management approach. Proceedings of the 39th Hawaii International Conference on Systems Sciences HICSS39. Kauai, Hawaii, 4–7 January. Chen, M., Cho, I-K., and Meyn, S. (2004). Reliability by design in distributed power transmission networks. University of Illinois, Urbana Champaign. Correljé, A.F. and. De Vries, L.J (this book). Hybrid electricity markets: The problem of explaining different patterns of restructuring. Chapter 2. Devine, W.D. Jr. (1983). From shafts to wire: Historical perspective on electriﬁcation. J. of Econ. Hist., 43(2), 347–72. EPRI. (2002). Review of the Current Status of Power Market Reforms in the United States and Europe. Palo Alto, CA. Federal Energy Regulatory Commission (2002). Standard Market Design Proposed Rulemaking. Washington, DC. Hunt, S. (2002). Making Competition Work in Electricity. Wiley. Joskow, P. (1997). Restructuring, competition and regulatory reform in the U.S. electricity sector. J. of Econ. Pers., 11(3), 119–38. Joskow, P. and Schmalensee, R. (1983). Markets for Power: An Analysis of Electrical Utility Deregulation. MIT Press. Michaels, R. (2006). Vertical integration and the restructuring of the U.S. electricity industry. Pol. Analy., No. 572, 1–32, published by the Cato Institute, 13 July. Newbery, D. (2006). Electricity liberalization in Britain and the evolution of market design. In Electricity Market Reform: An International Perspective (F.P. Shioshansi and W. Pfaffenberger, eds). Elsevier. O’Neill, R., Helman, U., Hobbs, B., and Baldick, R. (2006). Independent system operators in the USA: History, lessons learned, and prospects. In Electricity Market Reform: An International Perspective (F.P. Shioshansi and W. Pfaffenberger, eds). Elsevier. Rudnick, H., Barroso, L.A, Skerk, C., and Blanco, A. (2005). South American reform lessons – twenty years of restructuring and reform in Argentina, Brazil, and Chile., Pow. and Energ. Mag., IEEE, 3(4), 49–59. Sherman, R. (1985). The Averch and Johnson analysis of public utility regulation twenty years later. Rev. of Ind. Org., 2, 178–93. Sweeney, J. (2006). The California electricity restructuring, the crisis, and its aftermath. In Electricity Market Reform: An International Perspective (F.P. Shioshansi and W. Pfaffenberger eds.). Elsevier. Williamson, O. (1975). Markets and Hierarchies: Analysis and Antitrust Implications. The Free Press. Williamson, O. (1985). The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. The Free Press. Wilson, R. (1993). Nonlinear Pricing. Oxford University Press. Wilson, R. (2002). Architecture of power markets. Econometrica, 70, 1299–340. Zarnikau, J. (this book). Demand participation and demand response: Evidence from US ISOs, Chapter 9.

EBL

Chapter 2 Hybrid Electricity Markets: The Problem of Explaining Different Patterns of Restructuring A.F. CORRELJÉ AND L.J. DE VRIES Delft University of Technology, The Netherlands

Summary

EBL

This chapter explains the divergence in the design and structure of liberalized electricity markets. The main question is why there are so many “hybrid” markets, markets that are somewhere between their former state of a regulated monopoly and perfect competition. Electricity market design is found to be shaped by the relevant policy goals, the situation in the market at the outset of the restructuring process, and a variety of exogenous factors, outside the control of market participants and government policymakers. Moreover, feedback from market performance to the policy process is limited by time lags and bounded rationality. 2.1. Introduction Many electricity markets around the world have been or are in the process of being restructured with the purpose of introducing or expanding competition, but few have actually reached a state that could be described as commensurate with the economics textbook ideal of a liberalized, competitive market (cf. Joskow, 1996; Stoft, 2002). Perhaps the electricity markets in the United Kingdom, Argentina, Texas, New Zealand, Chile, and Alberta have come closest to this ideal, at least in terms of market design. However, they also have many idiosyncrasies, vestiges of their pre-liberalized state or results of political compromises, such as publicly owned competitive companies, preferred treatment of incumbents, inadequate network regulation, price regulations, or limited competition. Most electricity markets are still further removed from the textbook ideal and are somewhere in between their former pre-liberalized state and retail competition (Joskow, 2006; Newbery, 2005a,b; Victor, 2006; Rudnick et al., 2005; F. Sioshansi, this volume). If a function can be performed adequately by competitive ﬁrms, there is no need for public ownership. Nevertheless, public ownership remains common in restructured power 65

66

Competitive Electricity Markets

markets, so the state can often retain a strategic position. For instance, in many developing countries the single buyer remains a public entity. If the single buyer also owns some of the generating capacity, this leads to a situation in which it competes with independent power producers (IPPs) (Dehdashti, 2004). In many OECD (Organization for Economic Cooperation and Development) countries, the state, provinces, or municipalities maintain their shares in power production and networks. Another phenomenon is that markets have been restructured with all the features of a competitive wholesale market, but continue to be dominated by a single party – often, the former monopolist. Examples are Electricité de France, Electrabel in Belgium, or CFE in Mexico. Whereas ENEL, in Italy, was forced to divest a substantial amount of its generating capacity a few years ago, currently there appears to be a tendency in Europe to create “national champions” – large power companies that are expected to protect national interests in an increasingly international power market (Thomas, 2003; Haas et al. 2006). Finally, there are countries in which many of the requirements for a competitive market are met, both with respect to market design and industry structure, but in which the government continues to intervene beyond the minimal requirements for facilitating competition such as network regulation. Examples of government intervention are the use of a capacity mechanism, which inﬂuences investment in generating capacity (e.g., in Ireland, Spain, and in several markets in the United States) and wholesale or consumer price regulation. Price controls are common both in developing countries and in restructured OECD markets. Whereas in some markets a state of incomplete liberalization is clearly a transition phase, other markets appear static, stuck somewhere on the path to reform, while in a third group of markets the reforms are being reconsidered. Is this an indication that there are multiple steady states possible? Is there not a natural tendency for the design of markets to converge over time? Are markets typically moving toward more competition, or have some achieved a steady state? What drives these structural changes? What are the hurdles for further liberalization? What are the prospects for agreement upon a single market design, even within the European Union (EU) and the United States, where there are mechanisms available for coordinating state policies? The aim of this chapter is to investigate these “hybrid” markets and to provide an explanation of the apparent stagnation of the process. The main focus of this chapter is, How can the current divergence in the design and structure of liberalized electricity markets be explained and what are the prospects for convergence and agreement upon a single market design? In this chapter, countries that have undertaken attempts to introduce competition, from opening of their systems to new entry by IPPs to full-ﬂedged restructuring, but have not adhered to the textbook model for competition will be examined. Neoclassic economic theory provides the contours of the desired end state of a restructuring process by deﬁning the characteristics of an ideal market. It also provides a framework for diagnosing market imperfections (see Shuttleworth, 2000; Newbery, 2005b; Joskow, 2006; Jamasb and Pollitt, 2005; Haas et al., 2006; Kwoka, 2006). However, due to its prescriptive nature, neoclassic economic theory does not explain differences between the ways in which countries go about the restructuring process, or why markets in practice are so rarely fully organized according to the neoclassical economic textbook ideal. Neither does the neoclassic economic theory make clear why some companies are horizontally separated or privatized while others are not. There is a lack of understanding of the evolution of restructuring processes and hence also of their outcomes in as far as they do not conform to economists’ expectations. In this respect Heller and Victor (2004, p. 6) note: “Experts

EBL

Hybrid Electricity Markets

67

attribute this yawning gap between theory and practice to ‘politics,’ poor ‘rule of law’ and other ‘weak institutions’ that are needed to put the state on the sideline and to give the space for markets to operate. Yet, so far, nearly all scholarship has treated them as a residual category. [….] The lack of rigorous attention to these factors is particularly strange since political, legal, and institutional forces are hardly transient. Indeed, these factors appear to be the dominant ones in explaining the actual pace and character of market reforms in the electric power system in developing countries.” A promising approach for exploring the situation in these hybrid electricity markets is provided by Institutional Economics, a sub-discipline of economic science that deals with the evolution of institutional arrangements in markets (see, for example, North, 1990; Williamson, 1998; Glachant and Finon, 2000; Victor et al., 2006). In this theory, the behavior of market actors, as regards pricing, production, resource allocation, investment, strategies of horizontal and vertical integration, and so on, is assumed to be inﬂuenced by a market-speciﬁc body of rules and conventions: the institutional arrangements. Of course, the traditional economic perspective also relates the conduct of actors and their performance to structural characteristics of markets, like the degree of concentration, the shape of production functions, regulation. Yet, these characteristics are considered as a given, exogenous to a market. What distinguishes the approach of institutional economics is that it considers this body of rules as being an endogenous part of the market. Institutional arrangements are shaped by a path-dependent interaction between political, economic, and physical factors, driving deliberate interests, choices, and strategies of the policymakers, the ﬁrms in the industry, (groups of) consumers, and other parties involved, like international organizations and NGOs. This implies that an explanation of why certain types of hybrid markets have developed – and their future – should include the interaction between these factors and the resulting behavior of the actors in a speciﬁc market. The main elements in the institutional economy of power system restructuring are the situation at the start of the restructuring process, policy decisions regarding market design, related policies such as competition policy, fuel policy and environmental policy, and a number of exogenous factors that constrain the decision space, like the scale of systems and the availability of primary energy resources. Section 2.2 will further elaborate on this analytic framework. In Section 2.3 an overview is presented of the characteristics and achievements in respect of restructuring in a selection of electricity markets. The evaluation of the restructuring process of the electricity supply industry is based upon case studies of OECD countries, developing countries, and formerly communist countries (e.g., Heller and Victor, 2004; Victor and Heller, 2006; Newbery, 2005a; Joskow, 2005a; Rudnick et al., 2005; Sioshansi and Pfaffenberger, 2006). By considering North America and Europe as well as nonOECD countries, the full range of possible market designs becomes apparent. This section attempts to explain this diversity within the framework that was developed in Section 2.2. Section 2.4 provides an analysis of the country comparisons, while the conclusions are summarized in Section 2.5.

EBL

2.2. The Institutional Setting for Restructuring In this section an analytic framework will be developed. Section 2.2.1 starts with an analysis of the constraints that limit the decision space of governments as well as market players. Within this context, the motives of public authorities and the industry determine the direction of the restructuring process (Section 2.2.2) Section 2.2.3 describes the different market design variables. Given a speciﬁc restructuring policy, the strategies of

68

Competitive Electricity Markets

market parties determine the market outcomes in a dynamic process, as is described in Section 2.2.4. Section 2.2.5 summarizes the analytic framework. 2.2.1. The restructuring context The physical situation in a country provides a set of relatively “hard” constraints. A crucial element is the presence of indigenous energy resources, such as hydropower, coal, natural gas and oil, or – at the other end – dependence upon other countries for energy resources. Market size and degree of isolation matter too. Small isolated systems, in small countries or on islands, like Iceland, Malta, or in the Caribbean, cannot efﬁciently support multiple competing generating companies as a consequence of the relatively large minimum efﬁcient scale of the several types of generation units. This also impedes the use of speciﬁc technologies or fuels, such as large-scale coal plants. The geographic distribution of demand also plays a role: in thinly populated areas or small, remote concentrations of electricity demand it may be difﬁcult to create competition in supply (see also Weinmann and Bunn, 2004). A second set of constraints relates to macro-economic characteristics such as the level of economic development, the rate of demand growth, and the availability of investment capital. These factors inﬂuence the acceptability of changes in tariffs or prices to different categories of users, the need for investment, and ﬁnancing options for system expansion and/or rehabilitation. Three obvious categories of countries are, ﬁrstly, developing countries with a relatively stagnant economy; secondly, countries on the path of economic development and industrialization; and, thirdly, the OECD countries. The third category of constraints derives from the institutional and socio-political environment of a power system. North (1990), Williamson (1998), Glachant and Finon (2000), and Finon (2003) explain how informal institutions such as culture, traditions, and values affect the development of formal institutions, such as property rights, legislation, regulation, and the role of the (federal) state in the economy. In De Vries and Correljé (2006) it was discussed how formal institutions can be divided into general institutions, such as the polity, the judiciary, and the bureaucracy, and sector-speciﬁc institutions, such as sector legislation and regulation and jurisprudence, which are the main tools of market design. Arguing that the freedom of action for those who are in control of the reform process and the need to coordinate different aspects of the reform process are essential to the success of a restructuring process, Glachant and Finon (2000) consider the power of the central government a key factor with respect to the success of market reform. Table 2.1 provides an overview of the physical, macro-economic, and institutional constraints. Together they determine to a large extent the solution space that is available to governments that wish to restructure their power sectors. Within the context of these constraints, governments need to ﬁnd a balance between their own multiple objectives and those of the energy sector and of consumers. Some confusion may arise with respect to the difference between variables and constraints, as the starting value of a variable itself may pose a constraint on the liberalization process. For instance, the fuel mix at the outset of liberalization inﬂuences the economic and environmental performance of the sector for many years after the reforms have been instituted. However, this should not be considered a constraint, but rather the beginning value of the variable “fuel mix,” which may change over time. Undeniably, “take-off” variables such as the fuel mix have a signiﬁcant path-dependent impact, but analytically it is essential to make a clear distinction between ﬁxed constraints, as factors that lie outside the control of the actors within our system, and variables that are (more or less) under their control. Market design variables will be discussed in Section 2.3.

EBL

Hybrid Electricity Markets

69

Institutional factors

Economic factors

Physical factors

Table 2.1. Factors that determine the context of the restructuring process Factor

Impact

Natural endowment with energy sources

Presence or absence of primary energy sources drives the choice of primary fuels, the technical and economic characteristics of the sector, and interests and policies.

Physical size of the market

Due to scale effects, small markets are likely to be more concentrated. Larger markets may constitute a number of separate subsystems, with their own economic and institutional structure.

Geographic distribution of demand in relation to network capacity

Relatively dispersed demand and/or limited network capacity increase the likelihood of network congestion, which results in market fragmentation and limits competition.

Level of economic development and growth

Inﬂuences demand growth, the potential for investments, and institutional stability.

Growth rate of demand

Capacity investment lead times are long. With a high growth rate, large volumes of capacity must be under construction. Market signals or regulation must be effective. Stable demand, on the other hand, limits the “room” for new market entrants.

Financing options

Especially in developing or transition countries; with a weaker economy ﬁnancing options may be limited.

Ideology

General acceptability of and commitment to particular policies and institutions.

Institutional stability and rule of law

Facilitates investment and external funding; stabilizes and provides coherence in policies; helps align policy, regulation, and the legal framework.

Degree of institutional centralization and homogeneity

The power of the central government inﬂuences the coherence of policies and their support in terms of regional, sectoral, and social dispersion.

Inﬂuence of stakeholders

Strong stakeholders may be able to inﬂuence the reforms in their own interest.

EBL

2.2.2. Motives for restructuring A large number of states have embarked on the process of power sector restructuring, each starting from its particular national context (see Jamasb, 2002). A wide variety of motives can be observed for these attempts. Prior to liberalization, the electricity sector was considered a natural monopoly which, in the prevailing neoclassical approach, justiﬁed state intervention (Scherer, 1980; Stiglitz, 1986). In the United States, privately owned utilities where regulated by sector-speciﬁc Federal and State agencies. In Europe and its many (former) colonies and Latin America, public ownership became the dominant mode. In the communist world, of course, the state was

70

Competitive Electricity Markets

the owner and operator of virtually the entire economy. In addition to providing electricity at reasonable rates, the power sector policy served a number of public interests associated with these services, such as issues of safety, security of supply, acceptable prices for speciﬁc types of users, objectives of local and sector development, the supply of jobs, and – more recently – sustainability and environmental protection (see, for example, Foreman-Peck and Milward, 1994; Correljé et al., 2003). By the early 1980s, this perspective was replaced by the kind of “liberalism” that was associated with the ideas promoted by the governments of the late Ronald Reagan and Margaret Thatcher, which were based on monetarist and public choice theories. Efﬁciency, economic reform, and political power were sought through a reduction of taxes, “rolling back the state,” and by bringing market-driven competition into so-called “gold-plated” industries. Competition – modeled after the revised economic textbooks – was to be imposed upon public sectors wherever possible (Friedman, 1962; Demsetz, 1968; Helm, 2003; Parker, 2000). Gradually – initially only in a number of Anglo-Saxon countries and Chile – privatization and competition were introduced as the basic elements of structural change in the energy sector. Encouraged by the apparent successes of the early efforts, the introduction of competition became a common goal for many markets (see also Chao et al., Chapter 1 in this volume). As the competitive market paradigm became accepted in OECD countries, international organizations such as the World Bank and the IMF started to require developing countries to implement similar market reforms if they wished to be eligible for support. Thus, for many developing countries – as for some EU countries – reform was imposed, rather than a voluntary effort. Another possible reason for reform in developing countries is to attract private capital when the public owners cannot provide sufﬁcient investment. In addition, other policy objectives often play a role, such as social and economic stability, fuel policy, environmental policy, and CO2 trading. These policies may impact restructuring policy by excluding or enforcing particular fuel mixes and through systems of levies and subsidies (see Glachant and Lévêque, 2005). Moreover, competition policy may inﬂuence the success of restructuring. In general terms, it appears that the commitment to, and faith in, competition as a means of maximizing welfare is an important factor. The success of market reforms in other utility sectors may turn out to be an important example to policymakers (see F. Sioshansi’s Introduction).

EBL

2.2.3. Market design The preceding section provided an overview of the motives for power sector restructuring. Now the analysis will turn to the options for achieving these goals, the market design variables. Basic choices are the pace of market opening, the degree in which competition is introduced, the market model, decisions to restructure horizontally or vertically, and decisions to privatize (Glachant and Finon, 2000; Newbery, 2005a; Littlechild, 2003). An important means of achieving competition is by redesigning the market to make it attractive enough for new entrants to whittle away the market power of the incumbents. In designing markets, however, a tradeoff is to be made between competition and the investment climate. Ideally, the combination of economic and institutional structures and contract forms in a market reﬂects the risk structure in that market, involving market, fuel, regulatory, technical, and other risk components, as regards both short-term operating decisions and long-term investments (Joskow, 2005b; Alexander and Harris, 2005). The ﬁrst question is to what degree competition is to be introduced. A relatively light form of restructuring of publicly owned industries is their corporatization. State-owned

Hybrid Electricity Markets

71

enterprises normally operate under the responsibility and control of a particular department with “soft” budget constraints. Placing the enterprise at arm’s length from the responsible public body, while imposing a hard budget constraint, is a ﬁrst step toward enhancing the efﬁciency of the electricity industry and making it less responsive to political and interest group capture. A next step, introducing some competition without structurally addressing the state monopoly, is to allow IPPs. They sell to the state-owned monopolist (called the single buyer), who often continues to generate electricity (Heller and Victor, 2004). While this allows the state monopoly to dominate the system, the IPPs may provide benchmarks for performance and may increase their inﬂuence over time. The following step is to remove the single buyer and create a wholesale market in which there is wholesale competition between a number of generators, supplying large customers and distribution companies with a retail franchise. The ﬁnal step is the introduction of competition at the level of retail customers. To an increasing extent, these categories of market design allow for competition (Hunt and Shuttleworth, 1996). It is possible to open up the market to full retail competition at once, but it is also possible to gradually introduce competition by moving through these different models. The pace of restructuring is therefore a variable. Retail competition was often regarded as the desired end state of restructuring (it is, for instance, required by the EU in Directive 2003/54/EC), but in Chapter 1 of this volume Chao et al. make a case against competition at the retail level. Newbery (2002) also pointed out the advantages of retaining the retail franchise, such as retail companies being able to enter into long-term contracts for generation. With respect to the wholesale market model, there is a fundamental choice between integrated markets and decentralized markets (Hunt, 2002). Integrated markets are preferred in the United States. In these markets the system operator operates a mandatory pool, in which the physical and economic aspects of electricity trade are strongly connected. In decentralized markets, the preferred model in Europe, the system operator only has a technical function and supply and demand meet elsewhere, either bilaterally or in voluntary power exchanges. In the case of restructuring publicly owned utilities, privatization becomes an issue, as it is generally not considered necessary or desirable for government to be involved in competitive activities. Nevertheless, it is not uncommon for states to have their publicly owned companies operating in competitive electricity markets, perhaps due to the strong public goals with respect to this sector. Public authorities may engage private capital via a system of concessions, a ﬂotation of packets of shares in the industry on the exchange, or via a merger with a foreign ﬁrm. The choice of approach has important consequences for the resulting vertical and horizontal structure of the electricity sector and for the behavior of ﬁrms. The main variables are, ﬁrstly, the degree of privatization of the industry, ranging from a government monopoly via the admission of new privately ﬁnanced ﬁrms into the sector to full privatization. Secondly, there is the issue of what segments may be privatized – generation, wholesale and retail trade, the networks? In some cases, as in the Netherlands, it is argued that the networks, as critical infrastructures, should remain in public ownership (see Künneke and Fens, 2007). The third element involves the character of the new private owners: will they be anonymous shareholders, national or foreign ﬁnancing institutions, foreign electricity companies? A ﬁnal aspect involves the timing of privatization, relative to other elements of a restructuring. Depending on these choices, a new market structure emerges with a speciﬁc pattern of interests, strategies, and behavior. Obviously, the degree of privatization not only impacts the economic behavior of the ﬁrms in the industry, but also their activities in inﬂuencing

EBL

Competitive Electricity Markets

72

public authorities and regulators. Indeed, shifts in the structure of ownership bring about different constellations of interests, involving new parties and new strategies vis-à-vis the government, the regulators, and the market. The arguments for choosing a particular approach emerge only to a certain extent from pure efﬁciency considerations associated with restructuring. Other policy motives may include the size of the revenues to the state from privatizing the industry, the degree of control a government wishes to maintain over speciﬁc segments of the industry, international relations, security of supply issues, and the management of national energy resources such as gas or coal, and so on. If the goal of restructuring is to introduce competition, it may be necessary to break up existing companies into smaller ones or to force the incumbent to divest some of its assets. Indeed, a debate is currently taking place within the EU about the question as to what extent large cross-national ﬁrms should be allowed, as a means to enhance security of supply and investment potential, or whether these former national champions should be divided up in a number of potentially competing ﬁrms. Even if the market design is perfect, the potential beneﬁts of restructuring are not likely to be obtained if the market continues to be dominated by the former monopolist. Fringe competition may provide some beneﬁts, but due to economies of scale and the long life cycle of key assets such as power plants, it is not likely to lead to a level playing ﬁeld for competition very quickly. Therefore, horizontal unbundling is a key element of restructuring if multiple potential competitors are not already present. Public monopolies are in principle easier to split up than private companies (Finon, 2003). However, in case of public ownership, conﬂicts of interest with, for example, powerful labor unions may keep a government from acting. Some also consider it desirable to prevent or break up vertical integration between generation and retail. However, in Chapter 1 of this volume Chao et al. argue that this is a necessary instrument for risk hedging. This argument is supported by the observation that in most electricity markets there is a strong tendency toward vertical integration between generation and retail. The way the networks are regulated will fundamentally affect the network owners’ investment policies and therefore impact the adequacy with which demand growth and shifts in the pattern of supply and demand are met. Depending on the choice for a model, the issue of the position of the transmission and distribution networks’ monopoly becomes important. The regulation of network tariffs, the provision of regulated or negotiated third party access, and the degree of unbundling of the networks from competitive generation, trading, and retail activities strongly inﬂuence the overall effectiveness of competition in the market:

EBL

•

The quality of access regulation and the level and structure of tariffs affect the competitiveness of the supply of power and the development of trade. • The degree of unbundling also plays an important role in this respect, as it keeps incumbent owners of networks from obstructing access for new entrants and avoids cross-subsidization of competitive activities by non-competitive activities (see also Chapter 1). • Incentive regulation of transmission and distribution networks directly impacts the level of transport costs as a component of overall supply cost. In addition to access to transport, there are certain essential system operation functions that need to be provided. The main functions are scheduling and dispatch of transmission and distribution, balancing (in case of decentralized markets), congestion management, and ancillary services (such as black-start capacity and voltage control). These functions can be designed in multiple ways, but because of their relation with network management they are often provided by the transmission network manager (Knops, 2003).

Hybrid Electricity Markets

73

Trade between different electricity markets is often an issue. The connection with neighboring electricity systems may have a signiﬁcant impact upon the competitiveness of the market and the incentives to market parties. At the same time, it exposes the industries of connected countries to competitive forces, a phenomenon that may or may not be appreciated. In many markets, wholesale and/or retail prices are regulated. In this respect a balance needs to be struck between the interests of consumer groups, who often embody speciﬁc political power, and the incentives to the industry for providing sufﬁcient investment to cover future demand. It is still debated whether competitive electricity markets without price restrictions provide sufﬁcient investment incentives, or whether consumers would be better off with some kind of capacity mechanism. If there are price restrictions, theoretically a capacity mechanism is necessary to compensate generating companies for the foregone revenues; otherwise, they will under-invest (Hogan, 2005; Stoft, 2002). A capacity mechanism may also be needed, or be beneﬁcial, for other reasons (De Vries, 2004). An overarching issue is the role and position of the regulatory function. At what level should the regulator be placed? Local, provincial, national, or even supranational regulatory bodies may exist. The choice has consequences for the relation between the regulator and the regulated industry and its independence from other parts of the public realm. It also affects the degree of detail, speciﬁcity, and generality in which regulatory problems can be solved. A second aspect is the balance between sector-speciﬁc regulation and the application of general competition law. Both approaches have an ex ante component, addressing structural sector characteristics, prescribing behavior, and evaluating plans for mergers and acquisitions, and an ex post component for the monitoring and mitigation of abuse of market power. Consequently, the position of the regulator(s) and competition authorities within the policymaking arena and vis-à-vis interest groups has consequences for the transition toward a competitive market and the design of that market and, ultimately, for the allocative and dynamic efﬁciency of the sector and the market outcomes in terms of efﬁciency and welfare. The organization of the electricity supply industry prior to the restructuring process affects a number of these decisions by providing default choices. The number of generating companies at the beginning of the restructuring process, the degree to which they are integrated with network companies, the ownership structure of these companies, generating and network capacity, the capacity of network links with neighboring systems, whether there already is a regulator (e.g., for the regulation of private utilities) determine the starting conditions of the restructuring process. These conditions are relevant, as will be seen later in this chapter, because the restructuring process is subject to path dependency.

EBL

2.2.4. The dynamics of restructuring The solution space of governments with respect to restructuring – the range within which governments can choose the market design variables that were discussed in Section 2.2.3 – is restricted by the situation at the outset of the restructuring process and by the “hard” constraints that were described in Section 2.2.1. Within the context of the different constraints, governments need to ﬁnd a balance between multiple objectives with respect to the energy sector. To a large extent the speciﬁc balance between the policy objectives is a function of a country’s socio-political context, involving ideology and the representation of interest groups such as speciﬁc categories of energy consumers, producers of indigenous energy resources and their staff, the citizens at large, and components of the central and local administrations (see also Heller et al., 2004).

74

Competitive Electricity Markets

The solution space of market parties is restricted by the market design decisions of government, as well as by the exogenous constraints that were described in Section 2.2.1. Within these limits, parties optimize their positions through the contracts and transactions that they enter into with their suppliers and customers. An important element in this respect is the facilitation of trade by exchanges and other trading places by providing more or less standardized contracts. To a large extent, these market institutions emerge from the initiative of actors in the market, reﬂecting the opportunities for trade within a given system. The market structure evolves over time through mergers, takeovers, and the establishment of joint ventures. This process accelerates during the process of restructuring, as a consequence of ﬁrms’ growing insights in the newly developing context and the associated changes in their roles and opportunities (see also Kwoka, 2006; Jamasb, 2002). These strategic decisions by the ﬁrms, ﬁnally, determine the space within which they make their operational decisions, which lead to the market outcomes (Williamson, 1998). Based upon the performance of the market with respect to the government’s policy goals – and responding to public perception – government may adjust its policies (Willman et al., 2003; Correljé, 2005). The feedback from market performance to government may be weak, however, as North (1990) points out. First, for some of the market design choices, there may not be sufﬁcient information to reach an unambiguous conclusion about which is the better one. For instance, as of now, the experience with liberalized markets is too limited to provide deﬁnitive evidence whether a market design without a capacity mechanism provides sufﬁcient incentives for investment in generating capacity. Similarly, there is no consensus about the optimal design of a capacity mechanism. Given that the lead time for new generating capacity is several years, the length of the business cycle in the electricity generation business will probably be longer than a decade. Adding to that, many markets started restructuring with excess generating capacity, and it is clear that current experience with competitive markets is relatively short. There simply is not sufﬁcient empirical data with respect to the issue of generation adequacy. Economic models cannot sufﬁciently capture the details of existing markets, generation parks, and oligopolistic industry strategies to provide an unambiguous answer to the question of generation adequacy. Therefore, for the time being, answers to the issue will remain somewhat speculative (see De Vries, 2004; Cavaliere et al., 2007; Weinmann 2007). A second reason, according to North (1990), may be that the models with which actors process information may not be adequate. This is true both of formal models, such as simulation models, and of the “mental” models of reality that policymakers have in their mind. Regardless of the type, a model is a simpliﬁcation of reality and therefore always highlights particular elements of the real world. Different national policy traditions and socio-political backgrounds equip policymakers with differing mental models. Of course, these models improve over time if there is coherent feedback from what is happening in the real-world power sector and the economy at large. However, if the process takes too long, or when feedback signals are too weak and diffuse or contradictory, it stalls. The mental models are not consolidated and adjusted and insights remain ambiguous and open to multiple interpretations. In a highly complex system such as a competitive electricity market, this may easily be the case. The limited feedback is the main cause of path dependency. If it is not clear whether a change constitutes an improvement, the change is not made. This may explain the differences in market design between the United States and Europe, where many believe their own model is the best choice. Starting conditions were quite

EBL

Hybrid Electricity Markets

75

different, with much of the industry in the United States privatized before the restructuring process. Perceptions are also different, with European countries emphasizing the need for “national champions,” among others, as a balance against the large national energy companies that supply Europe with natural gas or against foreign electricity suppliers that would give preference to supplying their home country, in case of problems. This focus on national (or state) “champions” is absent in the United States. Finally, the blocking power of key actors may be another important cause of path dependency. Even if government has a clear vision for its market reforms, as the EU does with respect to electricity market restructuring, key actors such as powerful EU member states or incumbent companies may prevent it from happening. 2.2.5. Conceptual framework Combining the above elements leads to the conceptual framework that is represented in Fig. 2.1. This ﬁgure summarizes the relations between the different factors that inﬂuence the development of electricity markets. The market design process takes place within an institutional context: the process is inﬂuenced by informal institutions such as culture and values and constrained by formal institutions such as international treaties and the constitution. This context inﬂuences the choice and relative weight of the policy goals, as well as the general effectiveness of policymaking and implementation, as a function of the “power” and legitimacy of the government (Williamson, 1998; North, 1990). The institutional context also manifests itself through decisions of the past, as the institutional context is relatively static. The policy goals set the direction for the restructuring process. The policy goals, and the relative weights attached to them, vary widely. Objectives in related policy areas, such as energy security of supply, environmental and employment policy (fuel choice), social policy (energy prices) may also affect electricity market policy. The way these policy

EBL

Institutional context

Related policies

Objectives of restructuring

– Security of supply – Environment – Social issues

– Liberal policy and ideology – Economic efficiency – Investment in electricity system – International market integration – Externally imposed restructuring

Past policies

Horizontal unbundling, competition policy

Starting conditions

– Time delay – Bounded rationality

Market design

Exogenous factors

Table 2

Table 1

Market parties’ strategies

Market outcome

Fig. 2.1. Conceptual framework.

Feedback:

76

Competitive Electricity Markets

Table 2.2. Market design variables Variable

Consequences

Degree of market opening

Corporatization of a state monopoly, single buyer, wholesale market and competition in the retail market allow for an increasing degree of competition, but involve increasing transaction costs and requirements as regards the economic and institutional structure.

Pace of market opening

Being a leader or a follower vis-à-vis neighboring countries/states, or higher (EU, USA Federal) policy.

Integrated versus decentralized market

Integrated markets with mandatory pools, reduce transaction costs, but combine both economic and physical control over the system in the hands of a single party, potentially facilitating governance.

Public versus private ownership

Public ownership provides a means for direct control but entails policy captivity, may impede effective regulation and limits ﬁnancial resources. Private ownership requires political and institutional stability and regulatory commitment.

Competition policy and horizontal unbundling

Inﬂuences the competitiveness in market segments, trading off economies of scale and scope.

Network unbundling

EBL

Unbundling inﬂuences the independence of network managers and their interest in providing equal conditions for all network users.

Network regulation of network tariffs and access conditions

Inﬂuence the conditions for competition and the pressure upon network managers to work efﬁciently.

Congestion management method

Affects trade opportunities between regions.

Arrangements with neighboring networks and interconnector congestion management

Market integration may enhance the competitiveness but may also cause higher prices in the exporting country.

Balancing mechanism (in decentralized markets)

Balancing mechanisms affect cost and revenues of type generation, especially for intermittent sources, and inﬂuence entry.

Wholesale and end-user price regulation

Protects consumers, at the expense of investment stimuli to the industry.

Capacity mechanism

Different types of capacity mechanisms exist in order to stimulate investment in capacity.

Position of regulator

Ex ante or ex post regulation.

goals are combined inﬂuences the design of the market. The actual design choices are summarized in Table 2.2. The restructuring process is further constrained by a number of exogenous factors, as are listed in Table 2.1. These factors not only affect the market design process as policymakers take them into account, but they also affect the behavior of market parties directly.

Hybrid Electricity Markets

77

For instance, difﬁculty with ﬁnancing capacity expansion in developing countries will stimulate a market design that reduces investment risk; the combination of the incentives provided by the market design and the ﬁnancing possibilities together determine investment behavior. As argued above, due to path dependency, the starting conditions inﬂuence the restructuring process. It is easier to introduce competition into a market with fragmented ownership than into one with a monopoly. If the monopoly is state owned, the government has the power (if not always the willingness) to break up this company. Similarly, past fuel policies will affect the market for decades after the beginning of restructuring, due to the long life cycle of generation assets. Within these constraints, market design and competition policy provide the tools for restructuring. Feedback exists in that government observes the performance of the market and responds by changing policies, also stimulated by involved interest groups. However, this feedback is characterized by long time delays, due to investment lead times, incomplete information, and incomplete understanding of the meaning of that information. Faster feedback takes place in the form of lobbying by market players, trying to inﬂuence competition policy and market design decisions. 2.3. Hybrid Markets and Patterns of Restructuring The objective of this section is to investigate the evolution of a number of restructured electricity markets and to examine their prospects for convergence toward a single market design. A review of recent case studies on power market restructuring in a number of countries provides an insight into the factors that led them to their current situation as hybrid power systems. First the situation in a number of developing countries is discussed, followed by an examination of OECD countries.

EBL

2.3.1. Developing countries The power sector of Chile, the ﬁrst example of liberalization in the 1980s, was plunged into crisis when Argentina curtailed its natural gas exports to Chile in 2004. The power system had become dependent upon natural gas from Argentina as a consequence of its border conﬂict with the alternative supplier Bolivia. Investors debated between coal, using LNG for natural gas plants, and banking on a return of the cheaper Argentine gas. As they could handle these risks, which were partly political, they turned to the government for leadership (Rudnick et al., 2005). Despite being pioneers worldwide, these South American countries almost “froze” the reforms and – partly due to their political instability – no substantial adjustments have been made to cope with the challenges that emerged in recent years (Arango et al., 2006a, b). Argentina was also among the ﬁrst countries to reform its power sector in the early 1990s. Initially it was a showcase with a highly competitive generation market, but the deep macro-economic and political crisis in 2001 and 2002 and the subsequent devaluation of the peso led the government to freeze consumer electricity prices below cost. This, in turn, created a need for credit support to generating companies and government involvement in generation expansion. Eventually, the government resorted to the establishment of a new public ﬁrm (Arango et al., 2006b; Rudnick et al., 2005). Brazil was also confronted with the need to adjust its power sector by the end of the 1980s. Until then, the centralized power system of Electrobras, largely based on hydropower, had been expanding signiﬁcantly. A macro-economic crisis led to the need

78

Competitive Electricity Markets

for foreign investment in the sector. However, tariff adjustments and the lack of a stable regulatory environment posed obstacles to private investors (De Araújo et al., 2006; De Araújo et al., Chapter 15 in this volume). The adoption of a price mechanism that protected the operators of hydropower plants, at the expense of thermal gas-ﬁred power plants, further discouraged private investment. Reform efforts stalled during a major power crisis in 2001 and 2002, when a drought led to the rationing of electricity for 9 months. It was concluded that spot prices were inadequate for signaling the need for investment, considering the price volatility of a hydro-based system and the large ﬂuctuations in the growth rate of demand that were due to periodic economic crises. Within this context, the leftwing government that was elected in 2002 re-established central planning based upon hydro generation, involving cost-plus pricing and power purchase agreements for existing and new capacity, including a capacity mechanism in the form of mandatory forward contracts (Rudnick et al., 2005). In Mexico, the drive for reform has not been very strong, perhaps because the traditional system was successful in providing electricity to more than 96% of the population. Historically, the state-owned utilities could purchase oil at subsidized prices from the large publicly owned oil industry. In the mid-1980s, pressure arose to restructure the power sector, as part of a general wave of market reform in other sectors of the Mexican economy (Carreón-Rodriguez et al., 2003). This involved reforms in tariffs and fuel pricing, the creation of an IPP scheme by 1992, the creation of a (fragile) regulatory commission (CRE) by 1993, and the establishment of a new tariff structure in 2000. The current move toward populism, common in many countries in Latin America, has caused the policy to shift away from further private involvement in the power sector. Toward the end of the 1980s, the government of India was confronted with strong economic development, associated growth in power consumption, and limitations on public resources. The central government responded by trying to attract private investment in IPPs within the traditional framework of the State Electricity Boards (SEB), but this attempt failed. Consequently a decentralized approach was undertaken by a new reformoriented national government, in which the federal states were left free to adopt their solutions. In most states, the transmission company was given the role of single buyer. While this involved a combination of decentralization and corporatization, states were able to maintain their inﬂuence in the sector. The role of IPPs remained marginal. Political opposition to market-based tariffs by farmers and others who previously had enjoyed relatively low tariffs led to their continued cross-subsidization at the cost of industrial and commercial consumers (Tongia, 2003). As a consequence, industrial users started to engage in self-supply, thereby reducing their contribution to the cost recovery of the network and public generation. The ﬁnancial situation of the single buyers continued to be weak as a consequence. At the instigation of the central government, State Electric Regulatory Commissions were established, but they were unable to improve the situation. Currently there appears little motivation for reforms, but the gap between the regulated tariffs and the cost of investment may force change in the future (Tongia, 2003, p. 58). Several countries in South-East Asia, like Thailand, the Philippines, Malaysia, and Indonesia, have implemented modest forms of restructuring, generally involving IPP investments under power purchasing contracts. A crucial element was the robustness of these contracts during the Asian crisis at the end of the 1990s. Due to the relatively stable institutional and political framework, IPP investments in Thailand and the Philippines survived the South-East Asian crisis better than in Malaysia and Indonesia, which were affected by political clientelism, under which local and national politicians provided advantages and protection to the incumbent ﬁrms (Henisz and Zelner, 2001).

EBL

Hybrid Electricity Markets

79

The following picture emerges with respect to developing countries with strong demand growth. Reform was typically triggered by a large need for capital to ﬁnance the rapid expansion of the sector. For political reasons, the strong position of the publicly owned incumbents was often used for continuing the dominance of the state, which led to hybrid markets in which private initiative had a limited role. Usually, market restructuring was limited to allowing IPPs and/or establishing a single buyer model in order to minimize investment risk (Heller and Victor, 2004; Victor and Heller, 2006). In markets with wholesale competition, or a single buyer without retail competition, it is not necessary to unbundle the distribution networks and to regulate access to them. This model yields an ambiguous, unstable situation, particularly in the context of the macro-economic shocks that have occurred in many of these countries. Often, governments have sought to provide security to investors, for instance through a single buyer who would offer very long-term power purchasing contracts. Due to these macro-economic shocks, exchange risk became a serious issue to foreign investors that required adjustment of these contracts, which sometimes happened and sometimes did not. In most developing countries the introduction of retail competition has not been considered an option, as domestic consumption is often so low that it is not worth the restructuring and transaction costs. Especially in countries with substantial poverty, governments have attempted to protect the poor from high energy prices, for example by keeping consumer prices below average cost via cross-subsidization, or by squeezing the revenues of the IPPs. This of course discourages investment and causes a return to the initial problem of capital scarcity (see also Williams and Ghanadan, 2006).

EBL

2.3.2. OECD countries: the United States

The restructuring of electricity markets in the United States took place in a more decentralized manner than in the EU. A series of initiatives of the Federal Energy Regulatory Commission (FERC) brought about considerable development in open access to large-scale regional transmission systems, but FERC has not been able to implement a single market design. An explicit attempt at doing so in 2002, FERC’s Standard Market Design, failed due to opposition from the states and lukewarm support from the Federal Government. Consequently, the states have considerable room for deciding whether and how to restructure. Due to signiﬁcant differences between the electricity market context and between the interests of stakeholders in the states, a variety of restructuring paths have developed. According to Joskow (2005a), a key obstacle to the “difﬁcult and contentious” reform process in the United States is a lack of political willingness. A number of states did not go further than opening up their transmission lines to wholesale “wheeling.” A lack of network capacity and of network unbundling hampers wholesale trade in many regions. Other areas have better-developed wholesale markets, most notably the East Coast, California, and Texas. Market designs vary, but there is a preference for integrated markets with locational marginal pricing and for capacity mechanisms for stimulating investment in generating capacity (except in Texas). In addition to differences in the design of these capacity mechanisms, electricity markets in the United States vary with respect to price caps, retail price regulation and default supplier regulations, the position of the system operator, and network unbundling and access regulations. Differences between market designs reﬂect differences in interests and political inﬂuence of different stakeholder groups such as generators and consumers (cf. Cronin and Motluk 2006). The success of retail competition has been limited (Blumsack et al., 2005). The interest of consumers in switching to a different electricity provider has been minimal – perhaps

80

Competitive Electricity Markets

due to the attractiveness of staying with the regulated prices of “default providers” – and in general there was no noticeable decrease of retail prices (cf. Joskow, 2006; Newbery, 2005b). An interesting difference between the United States and Europe is the preference for integrated markets in the former and decentralized markets in the latter, using Hunt’s (2002) terminology. In Europe, there was a deliberate choice against the integrated model, as it would put too much power in the hands of the integrated system and market operator. In addition, the decentralized system, with physical bilateral trade, was deemed more transparent, while in the United States, the same reason was used in favor of integrated markets (Blumsack et al., 2005). This illustrates the degree to which bounded rationality plays a role. It can be concluded that there is a patchwork of more or less interconnected hybrid markets and supply systems in the United States. This patchwork is the consequence of the large variety in political decision making regarding the reform approach in the individual states, reﬂecting local conditions, as regards primary fuel supply, age of plant, structure of demand, and interconnection (see Joskow, 2005a). Many of the substantive issues of market design are decided at the state level, whereas the FERC provides general guidelines on particular aspects of interstate aspects of the market. Political and interest group pressure drives the states’ choices, primarily, toward optimizing the design of their local market from their own perspective. In the restructuring process in the United States, state governments have had to confront a number of regulated private companies. Some of these ﬁrms defended their positions, but others saw a window of opportunity for expanding outside their franchise areas. Nevertheless, unbundling typically involved a struggle over compensation payments for “stranded cost.” Incumbent ﬁrms tried to maintain their strong market position by obstructing competition or through mergers, if they were not taken over. In some states the process of restructuring was effective in the end, whereas in others it was limited to minimal compliance with FERC standards. Despite this mixed picture, the United States tends to be more inclined to pursue effective competition with a larger number of smaller players than European countries, perhaps because the United States is endowed with sufﬁcient indigenous energy sources whereas the EU is severely dependent upon imports of fossil fuels. For the latter, the purchasing power of their energy companies in the international fuel markets is thought to play an important role with respect to security of supply issues. A tendency toward large, national gas and power companies can be observed in Europe, whereas the North American gas industry is much more fragmented and hence more competitive. The US experience underlines the signiﬁcance of path dependency and learning effects. The industry structure at the outset of liberalization drove the political commitment to restructuring in particular states. States developed a variety of approaches to restructuring, based upon their particular circumstances, starting situations, and preferences. Over time, states have the opportunity to learn from each other, so something of a natural selection model develops. The success of some states in overcoming problems common to other states provides an example to others. The PJM system appears to have this role. PJM Interconnection is the regional transmission organization (RTO) that coordinates the movement of wholesale electricity in all or parts of Delaware, Illinois, Indiana, Kentucky, Maryland, Michigan, New Jersey, North Carolina, Ohio, Pennsylvania, Tennessee, Virginia, West Virginia, and the District of Columbia. In this respect Joskow’s argument (2006) about the lack of a generic model, as in the EU, may also be construed as an advantage. In the EU, the pressure of such a common “one size ﬁts nobody” model appears to generate

EBL

Hybrid Electricity Markets

81

considerable resistance, whereas the American diversity of approaches may in the end lead to voluntary convergence upon a preferred model (Weinmann, 2007).

2.3.3. OECD countries: Europe With the exception of the United Kingdom, the Netherlands, and Scandinavian countries, the restructuring of electricity markets in Europe is driven by EU policy. The European Commission (EC) started with a rather loosely deﬁned model of sector reform – with clear Anglo-Saxon principles – but has imposed increasingly strict requirements with respect to market design (CEC, 1996, 2003). The intention of the EU was to increase economic efﬁciency through the introduction of competition and thereby reduce end-user prices. With respect to the goal of sustainability, restructuring is also intended to facilitate the application of economic instruments, most notably the CO2 emissions trading system. The goal of security of supply, the third general policy goal, is expected to be served by restructuring, as the unbundling of the industry and the introduction of competition where possible should lead to optimal investment in each link of the value chain. Moreover, well-functioning markets should promote diversity, which provides resilience in case of disturbances. Recent reports by the EC show that these objectives have not been reached (CEC, 2006a, 2007b). The Sector Inquiry provides a thorough analysis, supported by a wealth of empirical data, that shows that European electricity markets are far from integrated and in most cases also far from competitive (CEC, 2006b, 2007c). A main problem is that the restructuring effort focused on market design but ignored competition issues and horizontal integration in most member states. Consequently, many member states’ electricity markets are dominated by a very small number of companies. Other important shortcomings are as given below:

EBL

•

•

• • • •

Insufﬁcient unbundling of the transmission networks from generating companies. The EC now strives for ownership unbundling or, if that is not feasible, the creation of independent system operators (ISOs). A lack of effective regulation (in particular in case of incompletely unbundled networks) and too strong a focus by regulators on their national markets instead of on the development of the EU internal electricity market. A continuing lack of sufﬁcient transparency. The availability of information varies signiﬁcantly between member states. Inadequate capacity of the infrastructure between member states. Insufﬁcient network security standards. Insufﬁcient signals for investment in generating capacity.

Many of the “hard” measures that are prescribed by the EC have been implemented only pro forma, which has not led to an effective single market or even a set of competitive national or regional markets. Member states strategically interpret and implement the directives and guidelines of the EU, their main concern being their own (future) position within the emerging EU energy market. Countries differ with respect to their commitment to, and faith in, competition, which appears to have an important impact upon the extent to which effective competition is created. Table 2.3 presents an overview of the present state of restructuring in the EU. Three groups of countries can be discerned

Table 2.3. The state of restructuring in the EU Wholesale market Concentration

GDP in PPS 2006

Population (in millions)

Greece Cyprus (d) Malta (d) Latvia Estonia France Belgium Slovenia Ireland Luxemb. (d)

85 88 70 52 65 107 118 84 139 257

11 0 7 0 4 2 3 1 3 59 9 10 4 2 4 0 5

100 100 100 95 90 85 85 70 85 14

100 100 100 100 100 95 95 95 90 ?

Public Public Public Public Public Public Private Public Public Mixed

62 35 0 76 10 70 90 75 56 57

Portugal Poland Slovakia Lithuania Spain Czech Rep. Italy Austria Germany Hungary Netherlands

70 51 59 55 98 76 99 123 110 64 126

10 5 38 2 5 4 3 4 42 3 10 2 57 9 8 1 82 5 10 1 16 2

50 15 75 50 40 65 55 45 30 30 25

80 35 85 80 80 75 75 75 70 65 80

Mixed Public Mixed Public Mixed Mixed Mixed Public Mixed Mixed Mixed

100 52 66 0 100 100 79 100 100 67 100

UK Sweden Denmark Finland Norway

117 116 122 113 169

59 6 8 9 5 4 5 2 4 6

20 15 15 15 15

40 40 40 40 40

Private Private Mixed Mixed Public

100 100 100 100 100

Largest Top 3 Ownership

Market opening

Sources: CEC (2005), (2006a,b), (2007a); Haas et al. (2006); EFTA 2007.

Transmission

Distribution

TSO Institutions Unbundling

DSO Ownership Unbundling

End-user price Ownership regulation

Regulator

Pool Bilateral None None None OTC/PX None None VIPP None

Legal Legal None Man. Legal Legal Own. Legal Own. Legal

Public Public n.a. Public Public Public Private Pub? Public Private

Acc. Legal Int. Int. Exempt Own. Legal None Own. Legal

Public Public Public Public Public Pub/loc Private Public Public Private

All All All All All All Dom. Dom. All Dom.

OK Weak Multi OK Weak Weak Weak Weak OK Weak

Bilateral Bil./PX Bilateral Bilateral Pool None PX None? Small Bilateral PX

Own. None Legal Legal Own. Own. Own. Legal Legal Legal Own.

Mixed Public Mixed Public Public Public Mixed Public Mixed Public Public

Legal None Legal Own. Legal Acc. Legal Legal Acc. Acc. Legal

Public Public Mixed Public Mixed Mixed Mixed Public Mixed Mixed Mixed

All All Dom. All All None All None Dom., com All None

OK Weak OK OK Weak Weak Strong Weak Weak Weak Strong

Large PX PX PX PX

Own. Own. Own. Own. Own.

Private Private Public Mixed Public

Own. None Exempt None Legal

Private Public Public Public Public

None None All Dom. None

Strong Strong Strong OK OK

EBL

Hybrid Electricity Markets

83

with more or less comparable characteristics regarding the implementation of the EU Directives. The ﬁrst group is characterized by a highly concentrated market and a near monopoly of the incumbent. This is generally paired with weak unbundling, absence of retail competition, and regulated tariffs for all consumers. This group includes many of the small new accession countries like in Estonia, Latvia, and Slovenia, which have low incomes and an underdeveloped post-communist institutional framework, and islands such as Malta, Cyprus, and Ireland, whose size limits the options for competition. The limitations of small, isolated systems are recognized by the EU and they may obtain a derogation of substantial parts of the EU Directive 2003/54/EC, art 26. Also traditionally state-oriented, monopolistic countries such as Greece, Portugal, France, and Belgium are in this group, in which a negative attitude toward competitive power markets dominates. Consequently, the governments in these countries see no pressing motivation, or possibility, for bringing about change. This is either because of their difﬁcult economic position or because of their being captive to electricity industry lobbies, via the channels of business and labor interest. This position is also reﬂected in the generally weak position of regulators (see Glachant and Finon, 2005; Finon, 2003; Meritet, 2007). In addition, the absence of indigenous resources has given rise to the nuclear energy policies of France and Belgium. In the second group the market is moderately concentrated. Most countries have opened their markets fully, but the degree of network unbundling varies and in some cases price regulation is applied. This group is fairly heterogeneous and includes new accession countries such as Hungary, the Czech Republic, Poland, and Lithuania, plus long-standing member states such as Germany, Spain, Austria, Italy, and the Netherlands. These countries are characterized by the fact that their electricity industry always included multiple companies. In the central European countries the common objective to get rid of their past dependence on energy supply from Russia is a strong driver for change. Particularly these industrialized countries may harvest the advantages of improving the competitiveness of their economies. However, social and regional antagonisms need to be solved and economic problems arise in this process. The group of long-time EU members is characterized by a variety of governance traditions, ranging from corporatism to clientelism. Generally, intra-industry relations are relatively informal, but the industry is at some distance from the state. These countries all depend to a large extent on imported fuels. Generally, public policy in these countries needs to balance between opening up the power sector and their strategic interest with respect to security of supply and the survival of their national industries. In Spain and Germany, the use of indigenous coal and lignite has been supported to protect the jobs of miners. The multiple objectives cause conﬂicts between the state, regulators, and the industry, while political parties remain divided over restructuring issues. Compromises in the restructuring policy are the consequence. For example, Germany was one of the ﬁrst EU members to fully open up its electricity market. However, as the cooperating incumbents were not required to unbundle their networks and because network access and tariffs were not regulated, the market remained virtually closed to new entrants. Only recently the process was given an impulse by the establishment of an independent regulator and the further implementation of the EU Directive (see Brunekreeft and Twelemann, 2005). During the 1990s, the Netherlands appeared to make a strong shift toward the Anglo-Saxon model. However, after years of intense conﬂicts between the electricity sector and government, there now appears to be a movement back toward a more consensus-based approach in which network tariff regulation is less strict and companies are more involved in rule making. As a defense against foreign inﬂuence in the industry via takeover, the Dutch government decided

EBL

84

Competitive Electricity Markets

to prohibit privatization of the networks (Van Damme, 2005). The Spanish process of adjustment is also rather conﬂictive. From its protectionist background, the Spanish state chose for a policy of restructuring from a position of “strength,” allowing the national industry to compete with foreign ﬁrms, at home and abroad. Given the diversity of interests in the Spanish system and their relations with the polity, this policy does not produce clear guidelines and choices. This is also illustrated by the ongoing struggle between the government and industrial interests over the position of the electricity company Endesa (Crampes and Fabra, 2005; Serrallés, 2006). The third group of countries, consisting of the United Kingdom and the Nordic countries, are the most liberal, with low market concentration, integrated markets (in Scandinavia), and retail competition. These countries have deliberately chosen, on an ideological basis, to liberalize their power systems, by dismantling their former monopolies and restructuring their industries. In the United Kingdom, an important additional objective for restructuring was the limitation of the coal miners’ union power. The strength of these countries appears to be their relative independence from imports and their access to a variety of fuels, which has ascertained their belief that the process of allocative decision making can be trusted to the “internal” market. Moreover, the institutional context of these countries places organized industrial interests at some distance from policymaking, while economic growth and ideology make them less vulnerable to the consequences of occasionally high consumer prices (Kinnunen, 2004). This process is illustrated by the political reactions to the electricity prices spikes in the Nordpool market in 2001 and 2003 and the price hikes in the UK gas market during the winter of 2005–06. An important explanatory factor for the market situation in European countries appears to be the initial ownership structure of the power sector. Apparently, the existence of national monopolies, as in France, Portugal, Belgium, and Greece, makes it difﬁcult for a government to create competition. The only exception is the United Kingdom that split up its Central Electricity Generating Board. The general solution is “fringe competition,” the facilitation of “competitive” entry alongside the monopoly. However, because of the continued alliance between the dominant public monopoly and the state, the development of new institutional arrangements and the regulatory body proceed slowly. Price regulation needs to be maintained against the market power of the monopoly, but this often harms the interests of (potential) new entrants (Glachant and Finon, 2005). If the government is confronted with a number of vertically integrated ﬁrms that are publicly owned, whether by the state or by provinces and/or municipalities, the situation appears relatively straightforward. Unbundling, the introduction of access rules and tariffs, and wholesale and retail competition can be established by political decision making. Main causes of resistance are conﬂicts of interest between public bodies at different levels and interest groups clamoring for consumer protection, environmental issues, and the unions. Eventually, this state of affairs may progress and lead toward privatization. When the government needs to confront a number of private ﬁrms the situation is more complex. In the EU it can be doubted whether effective unbundling is legally possible, as it may be considered expropriation. States that only reluctantly accepted liberalization may try to protect the interest of their “national” ﬁrms through inﬂuencing its quasi-independent regulator, assuming that they will serve the country better than new entrants. This argument runs two ways. If new entrants are private ﬁrms, they are thought to be driven only by proﬁts and ignorant of the public character of electricity supply. If the new entrants are foreign public ﬁrms, the situation is also suspect, as these ﬁrms are thought to give preference to their home market in case of supply problems.

EBL

Hybrid Electricity Markets

85

The European markets have witnessed the emergence of a set of large multinational conglomerates, the so-called seven brothers (see Thomas, 2003; Green, 2006; Haas et al., 2006) that extend their activities by taking over energy companies in their home country and abroad. These companies may be considered as (potential) competitors in the future pan-European market. Their purchasing power in the international fuel markets is often brought forward as an important asset with respect to security of supply issues, as it provides a counterbalance to the market power of Gazprom and the Middle Eastern gas suppliers. On the other hand, the EC recently announced plans to take on the dominant position of companies such as Electricité de France, E.On, RWI, and ENI (EC, 2006, 2007). In response, governments of countries that are not home to one of the seven brothers feel an even stronger need to protect their power market against “foreign domination,” either by establishing their own mini-champions or by keeping the market relatively closed to takeovers, for instance by a minimalist implementation of the Directives or by not privatizing their companies. So while these governments may sound supportive of the notion of introducing competition and may not oppose the corresponding changes to the market design, they may oppose attempts to really restructure the market. This situation causes inconsistencies, on/off policies, and a slow development of an effective market. It also leads to competition between the EU member states’ “national” companies, which has led to the reintroduction of a great deal of economic nationalism in the restructuring process, often veiled in terms of environmental protection, security of supply measures, or public service obligations.

2.4. Analysis

EBL

This section will review how the empirical evidence that was presented in Section 2.3 ﬁts with the conceptual framework that was proposed in Section 2.2.5. While in general, the empirical observations appear to match our expectations, it is now possible to ﬁll in the blanks in the conceptual framework. Clearly, restructuring processes are driven by a wide variety of objectives. Much of the literature describes projects that were meant to evaluate and support policies of particular organizations, such as the World Bank and regional organizations such as the EU, APEC (2000), and industrial associations, which is why these studies tend to focus on sets of relatively comparable countries with similar objectives. Without such a restricted focus, a rich variation in motives for restructuring and in the backgrounds of such processes emerges. After the adoption of the single European market in 1985, liberalization became a priority for the EC, initially as the main instrument for removing the existing intra-communal barriers to trade and later as an objective on its own merit (Haaland and Matláry, 1997). In addition to economic efﬁciency, an important motive is the idea that open markets would increase the interdependence of European countries and therefore bring them closer together. However, the goal of competition is not shared by all EU member states and some have restructured merely in order to comply with the Directives. Even when states believe in competition as a means of maximizing social welfare, they will need to balance the commitment to restructuring with other policy goals. Country-speciﬁc motives also play a role, such as in the case of Britain’s Prime Minister Thatcher, who used liberalization of the electricity industry and the subsequent dash for gas as a means to put an end to the political power of the coal unions (Parker, 2000). Similar decisions are

86

Competitive Electricity Markets

made by states in the United States that need to comply with FERC regulations (Joskow, 2006). The end of the Communist regimes in the Soviet Union (Tompson, 2004; Balaschak, 2006) and its former satellite states in Eastern Europe and elsewhere was the starting point of a “transition” toward a market-based economy. In these countries, restructuring the electricity sector was part of the general economic transition, which was partly an autonomous process and partly driven by the prospect of EU membership. Developing countries, which inherited the industry structures from their former colonizers, have announced restructuring plans on a large scale. In some cases restructuring is a direct corollary of the existing political ideology and economic strategy. However, Victor and Heller (2004) argue the principal driving force for reform is ﬁnancial: “Even in cases where reformers have held the reins of power, substantial reform has rarely followed until the state dominated system is bankrupt and the lights are dimming.” Within the context of a ﬁnancially fragile public sector, this meant that investments in the supply and provision of electricity would have to be undertaken mainly by private operators and funded by private funding agencies. In many developing countries, however, restructuring was imposed by the IMF or the World Bank as part of economic support packages (see Rosenzweig et al., 2004; Yi-chong, 2006; Potts Voll et al., 2006). Generally, these packages are geared toward providing the necessary investment in infrastructures and institution building to facilitate further development of these economies (Thomas, 2006). Hence, expanding and renovating the electricity system in developing countries requires a certain degree of restructuring and a shift in the role of government and publicly owned ﬁrms. The relatively weak capacity of the public organizations and the allegedly more efﬁcient operation of private sector ﬁrms is also a motivation for restructuring (Bacon and Besant-Jones, 2001; Victor, 2006). In other developing countries the situation is somewhere in between these two extremes. Restructuring may be undertaken voluntarily, not because of an inherent faith in competitive markets, but as a means of attracting foreign investment. The restructuring process in these countries is vulnerable. A shortage of electricity in a competitive market or high fuel prices (which may also be caused by a devaluation of the local currency) may cause high electricity prices, which may be socially and politically unacceptable. The temptation would be strong to re-regulate prices, which, without compensating measures, would discourage investment, returning the country to the original problem of attracting investment. So whereas the introduction of competition is often presented as a logical and inescapable consequence of the superiority of “the market” as an economic coordination device, it can be observed that a range of different motives exist in different countries and at different times. Shifts in drivers and process dynamics lead governments and market parties to adjust their strategies (see Green, 2006). Hence, when evaluating the progress in restructuring and the emergence of hybrid markets, it is necessary to distinguish the initial drivers behind the process from later adjustments to them. Indeed, Rudnick et al. (2005) and Sioshansi (forthcoming), respectively, refer to “second stage” and the “reform of the reforms.” The impact of related policies, such as environmental policies, upon market design appears to be limited. These policies are typically achieved through other means, such as standards or taxes. The main exception is security of supply of primary fuels, which may run counter to the drive to create a market with many small competitors, as the perception may exist that the market power of large ﬁrms is needed to secure good fuel import contracts.

EBL

Hybrid Electricity Markets

87

In summary, with respect to the motives for restructuring, empirical analysis shows that governments may or may not: • • •

•

•

•

support a neo-liberal program in order to enhance the efﬁciency of public sector management in the broadest sense by privatization and restructuring; seek to enhance the efﬁciency of their power systems speciﬁcally, to achieve lower supply prices, to improve quality of supply, and to diversify their fuels; seek private investment in their electricity industry, to facilitate an expansion of their supply potential and therefore accept some degree of restructuring, most likely involving higher prices to (groups of) consumers; seek to achieve market integration with (groups of) adjoining countries and accept some degree of reduction in their autonomy over “public interest” sectors, in exchange for advantages for other economic sectors; support integration with neighboring markets, for the purpose of exploiting (potential) advantages in exporting or importing power, or by engaging in the electricity industry of associated countries; restructure merely to satisfy requirements of other authorities, such as the EU, the FERC, the IMF, or the World Bank.

Of the exogenous factors that were proposed in Section 2.2.1, two, apart from the cultural and ideological commitment to competition as a means of maximizing welfare, stand out. The ﬁrst is the level of economic development, in which a high income per capita generally correlates with a lower rate of growth in electricity demand and the presence of a relatively stable legal and policy framework. Lower income countries often experience higher rates of growth in their electricity demand, but also experience higher risks of economic and institutional instability. Consequently, investment risk is higher and they face greater difﬁculty in attracting sufﬁcient investment capital. Indeed, a competitive energy-only market signiﬁcantly raises investment risks. To reduce this risk, the introduction of competition can be limited to wholesale competition or be accompanied with a capacity mechanism (De Vries, 2007). A second important exogenous factor is the availability of indigenous primary energy resources, like hydropower, coal, petroleum, or gas. Hydropower fundamentally changes the dynamics of the market, as peak capacity is cheap, but the total amount of energy stored in the reservoirs needs to be spread out over the dry season. Coal exploitation is often associated with employment and social issues. Petroleum production generally yields high revenues to the state, thus reducing ﬁnancial pressures to restructure, while supporting particular political interest coalitions in maintaining the status quo. The presence of (un)tapped natural gas reserves may support pressures to restructure the power industry, as it promises low (capital) cost and an efﬁcient expansion of generation capacity. Here geopolitics and the organization of the gas industry are important variables. Countries that are endowed with sufﬁcient indigenous energy sources show a stronger tendency to pursue effective competition with a number of smaller players. In contrast, in Europe, resistance to privatization and horizontal unbundling is sometimes linked to concerns about fuel import dependency, in particular with respect to natural gas (Correljé and Van der Linde, 2006). Due to path dependency, history matters. Decisions made in the past affect future options. Therefore, the situation at the start of the restructuring process has a signiﬁcant impact upon the restructuring process and its outcome. Most important is whether at the outset of restructuring there is a national monopoly in the power sector or a number of regional ﬁrms. Glachant and Finon (2000) consider the presence of an incumbent monopoly

EBL

88

Competitive Electricity Markets

the “Achilles’ heel” of restructuring as it will be difﬁcult for a government to split it up in order to create competition, especially if the company is privately owned. A likely outcome is the facilitation of “competitive” entry alongside the incumbent. Empirically, this situation is observed in developing countries, where IPPs were encouraged, and in some EU countries. If the alliance between the (public) monopoly and the state is continued, new institutional arrangements and a regulatory body develop only slowly. To mitigate the market power of the monopoly, price regulation needs to be maintained, but this often harms the interests of (potential) new entrants. Also, with respect to vertical unbundling, ownership is an important factor. If incumbent, vertically integrated ﬁrms are publicly owned, unbundling, the introduction of access rules and tariffs, and wholesale and retail competition are more or less easily established through the political process. In addition to opposition by the incumbents, the main sources of resistance to restructuring may be public bodies at different levels and with different interests, consumer protection interest groups, environmental advocacy groups, and, last but not least, trade unions. When a government wishes to restructure private ﬁrms, the situation is much more complex. Unbundling often involves a struggle over compensation payments for “stranded cost.” Moreover, the incumbent ﬁrms will try to maintain their oligopolistic position by obstruction of the market or by mergers, or they may be taken over by foreign ﬁrms. On the other hand, if the state only halfheartedly pursued liberalization in the ﬁrst place, it may try to protect the interests of the “national” ﬁrms, for instance because they are presumed to serve the country better than foreign parties, the more so if these ﬁrms have other public tasks. This situation causes inconsistencies, on/off policies, and a very slow development of an effective market. Taking all these issues into consideration, what is the prospect for agreement upon a single market design? To the extent that exogenous factors or fundamental differences in policy objectives are the causes of differences in the market design, they will likely have a lasting effect and multiple market designs are likely to continue to coexist. Examples are markets in which concerns for security of supply, particular physical conditions, or high investment risk have prompted government intervention. On the other hand, when variations in market design are due to differences in opinion regarding the means, in a context of similar policy goals and similar exogenous conditions, one would expect that mutual learning would eventually lead to convergence of the market design. North (1990), however, warns that this may be a slow process. The bounded rationality of the actors contributes to path dependency in which preexisting patterns are replicated with only modest alterations. This may explain, for instance, the preference for integrated markets (with mandatory pools) in the United States versus the decentralized markets of Europe, as well as the presence of capacity markets in the United States and their absence in Europe.

EBL

2.5. Conclusions This chapter has presented a framework for explaining the many differences in the approach to electricity market restructuring around the world. A number of policy objectives exist with regard to power market restructuring; the particular institutional context has a strong inﬂuence upon which policy objectives are chosen and how they are prioritized. Implementation of these objectives takes place through two main policy areas: the market design process and competition policy (including the policy with respect to horizontal unbundling of dominant incumbent ﬁrms). This process is constrained by external factors such as the economic climate and the physical situation in a country. In addition,

Hybrid Electricity Markets

89

the situation at the outset of the restructuring process inﬂuences the process itself. Due to path dependency, this inﬂuence may be long-lasting. Feedback exists in the sense that governments observe the performance of the electricity market and respond by adjusting their policies. This feedback is limited by time lags, incomplete data, bounded rationality, and the fact that much of the feedback represents the perspectives of lobbyists. The limitations of the information feedback loop give rise to path dependency: given uncertainty about the best direction for policy, decisions that generate the least resistance are favored, which means those that depart the least from the status quo. A second cause of path dependency may be that the government is not able to overcome resistance to change by vested interests. This framework has helped to explain the diversity of restructuring processes and market designs that can be observed. The prospects for future convergence of these different market designs hinge upon a number of factors: •

The policy goals and their relative weights vary. A key factor is whether the restructuring of a country’s power sector is ideologically and politically motivated or whether it is more or less forced upon a country by a higher authority, the latter often leading to a half-hearted and unstable approach. • Due to path dependency, the situation at the start of a process of restructuring signiﬁcantly inﬂuences the potential for competition, especially the degree of concentration and whether or not the industry is already privatized. • Exogenous factors, such as physical scale, endowment with natural resources, and macro-economic conditions, may have a dominant inﬂuence upon the design of the market. Countries that are endowed with sufﬁcient indigenous energy sources tend to be more inclined toward pursuing effective competition with a number of smaller players than countries that are severely dependent upon imports of fossil fuels.

EBL

When exogenous factors or fundamental differences in policy objectives are the causes of differences in the market design, there is little prospect for convergence of market designs. Otherwise, if the circumstances and policy goals are similar, mutual learning should lead to convergence, but this process is limited to similar markets and will be slow due to the imperfections of the policy feedback loop. References Alexander, I. and Harris, C. (2005). The regulation of investment in utilities: concepts and applications. World Bank Working Paper no. 52, The World Bank, Washington. APEC (2000). Electricity sector deregulation in the APEC region. Asia Paciﬁc Energy Research Center, March, Tokyo. Arango, S, Dyner, I., and Larsen, E.R. (2006a). Lessons from deregulation: Understanding electricity markets in South America. Uti. Pol., 14, 196–207. Arango, S, Dyner, I., and Larsen, E.R. (2006b). Understanding the Argentinean and Colombian electricity markets. In Electricity Market Reform, An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier, pp. 595–616. Bacon, R.W. and Besant-Jones, J. (2001). Global electric power reform, privatization, and liberalization of the electric power industry in developing countries. Ann. Rev. of Energ. and the Env., 26, 331–59. Balaschak, J. (2006). Restructuring the Power Industry in Russia, 15 December. Available at: http:// www.amcham.ru/doing_business/oil_and_gas_issues/restructuring_power_industry_in_russia. Blumsack, S.A., Apt, J., and Lave, L.B. (2005). Lessons from the failure of U.S. electricity restructuring. The Elec. J., 19(2), 5–31. Brunekreeft, G. and Twelemann, S. (2005). Regulation, competition and investment in the German electricity market: Reg TP of REGTP. Energ. J., Special Issue, 99–126.

90

Competitive Electricity Markets

Cavaliere, C., Larsen, E.R., and Dyner, I. (2007). The privatization of EEB: From cash drain to major contributor. Energ. Pol., 35, 1884–95. Carreón-Rodriguez, V.G., Jiménez San Vicente, A., and Rosellón, J. (2003). The Mexican Electricity Sector: Economic, Legal and Political Issues. Working Paper #5. Program on Energy and Sustainable Development, Center for Environmental Science and Policy Stanford Institute for International Studies, November. CEC (1996). Directive 96/92/EC of the European Parliament and of the Council of 19 December 1996 concerning common rules for the internal market in electricity. Off. J. of the Eur. U. 1997, L 27, 20–9. CEC (2003). Directive 2003/54/EC of the European Parliament and of the Council of 26 June 2003, concerning common rules for the internal market in electricity and repealing Directive 96/92/EC. Of. J. of the Eur. U. 2003, L 176, 37–55. CEC (2005). Report on progress in creating the internal gas and electricity market. Brussels: CEC. CEC (2006a). Sector Inquiry under Art 17 Regulation 1/2003 on the gas and electricity markets (Preliminary report). Brussels: CEC. CEC (2006b). Sector Inquiry under Art. 17 of Regulation 1/2003 on the gas and electricity markets (ﬁnal report), SEC 1724. Brussels: CEC. CEC (2007a). Prospects for the internal gas and electricity market: Communication From The Commission To The European Council And The European Parliament (COM(2006) 841 ﬁnal). Brussels: CEC. CEC (2007b). An energy policy for Europe: Communication from the Commission to the European Council and the European Parliament (COM(2007) 1 ﬁnal). Brussels: CEC. CEC (2007c). DG competition report on energy sector inquiry. Brussels: CEC, DG Competition. Chao, H., Oren, S., and Wilson, R. (this book). Reevaluation of vertical integration and unbundling in restructured electricity markets. Chapter 1. Correljé, A. F. (2005). Dilemmas in network regulation: The Dutch gas industry. In Innovations in liberalized network industries: Between private initiatives and public interest (R. Künneke, J. Groenewegen, and A. Correljé, eds). Edward Elgar, 115–50. Correljé, A.F., Van der Linde, J.G., and Westerwoudt, T. (2003). Natural Gas in the Netherlands: From cooperation to competition? The Hague: Clingendael International Energy Programme/Oranje Nassau. Correljé, A.F. and Van der Linde, C. (2006). Energy supply security and geopolitics: A European perspective. Energ. Pol., 34, 532–43. Crampes, C. and Fabra, N. (2005). The Spanish electricity industry: Plus ça change The Energ. J., European Energy Liberalization Special, 127–53. Cronin, F.J. and Motluk, S.A. (2006). Reviewing electric distribution restructuring in Ontario: Policy without substance or commitment. Uti. Pol., 14, 1–7. De Araújo, J.L.R.H. (2006). The case of Brazil: Reform by trial and error? In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. De Araújo, J.L.R.H., Correia, T., Da Costa, A.M.d.A., and Melo, E. (this book). Reform of the reforms in Brazil: Problems and solutions. Chapter 15. De Vries, L.J. (2004). Securing the public interest in electricity generation markets, The myths of the invisible hand and the copper plate. Ph.D. dissertation, Delft University of Technology, Faculty of Technology, Policy and Management. De Vries, L.J. (2007). Generation adequacy: Helping the market do its job. Uti. Pol., Article in press. De Vries, L.J. and Correljé, A.F. (2006). Hybrid electricity markets. In Proceedings, 26th USAEE/IAEE North American Conference on Energy in a World of Changing Costs and Technologies, Ann Arbor, Michigan. Dehdashti, E.S. (2004). Developing countries; restructuring with beneﬁts from competition (or not). IEEE Pow. & Energ. Mag., September/October, 16–23. Demsetz, H. (1968). Why regulate utilities? J. Law and Econ., 11, 55–65. EFTA (European Free Trade Association) (2007). Energy sector enquiry: Concluding report. EFTA Surveillance Authority, 10 January. Finon, D. (2003). Introducing competition in the French electricity supply industry: The destabilisation of a public hierarchy in an open institutional environment. Working Paper. Cambridge: the Cambridge-MIT Institute.

EBL

Hybrid Electricity Markets

91

Foreman-Peck, J. and Milward, R. (1994). Public and Private Ownership of British Industry 1820-1990. Oxford: Clarendon Press. Friedman, M. (1962). Capitalism and Freedom. Chicago: Chicago University Press. Glachant, J.-M. and Finon, D. (2000). Why do the European Union’s electricity industries continue to differ? A new institutional analysis. In Institution, Contracts and Organizations: Perspectives from New Institutional Economics (C. Ménard, ed.). London: Edward Elgar. Glachant, J.-M. and Finon, D. (2005). A competitive fringe in the shadow of a state monopoly. The Energ. J., European Energy Liberalization Special, 181–204. Glachant, J.-M and Lévêque, F. (2005). Electricity internal market in the European Union: What to do next? Working Paper. SESSA research project on European electricity regulation, www.sessa.eu.com, September. Green, R. (2006). Electricity liberalisation in Europe – how competitive will it be? Energ. Pol. 34(16): 2532–41. Haaland Matláry, J. (1997). Energy Policy in the European Union. Houndmills: Macmillan. Haas, R., Glachant, J.-M., Keseric, N., and Perez, Y. (2006). Competition in the continental European electricity market: Despair or work in progress? In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Heller, T.C., Tjiong, H.I., and Victor, D.G. (2004). Electricity restructuring and the social contract. Working Paper #15. Program on Energy and Sustainable Development, Center for Environmental Science and Policy, Stanford University, Stanford, May. Heller, T.C. and Victor, D.G. (2004). A political economy of electric power market restructuring: Introduction to issues and expectations. Working Paper #1. Program on Energy and Sustainable Development, Center for Environmental Science and Policy, Stanford Institute for International Studies, Revised, May. Helm, D. (2003). Energy, the State and the Market: British Energy Policy since 1979. Oxford: Oxford University Press. Henisz, W.J. and Zelner, B.A. (2001). The political economy of private electricity provision in Southeast Asia, WP 2001—02. Working Paper of the Reginald H. Jones Center The Wharton School, University of Pennsylvania. http://jonescenter.wharton.upenn.edu/papers/2001/wp01-02.pdf Hogan, W. (2005). On an ‘Energy-Only’ Electricity Market Design for Resource Adequacy. Cambridge, MA: Harvard University. Available from http://ksghome.harvard.edu/∼whogan. Hunt, S. (2002). Making Competition Work in Electricity. New York: John Wiley & Sons, Inc. Hunt, S. and Shuttleworth, G. (1996). Competition and Choice in Electricity. Chichester: John Wiley & Sons, Inc. Jamasb, T. (2002) Reform and regulation of the electricity sectors in developing countries. DEA Working Paper WP 0226, University of Cambridge. Jamasb, T. and Pollitt, M. (2005). Electricity market reform in the European Union: Review of progress toward liberalization & integration. The Energ. J., European Energy Liberalization Special, 11–42. Joskow, P.L. (1996). Introducing competition into regulated network industries from hierarchies to markets in electricity. Ind. and Corp. Ch., 2, 341–82. Joskow P.L. (2005a). Markets for power in the United States, an interim assessment. The Energ. J., 27(1), 1–37. Joskow, P. L. (2005b). Vertical integration. In Handbook of New Institutional Economics (C. Menard and M. Shirley, eds). Springer. Joskow, P. (2006). Introduction to electricity sector liberalization: Lessons learned from cross-country studies. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Kinnunen, K. (2004). Electricity network regulation – practical implementation in the Nordic countries. In Trends in Infrastructure Regulation and Financing. International Experience and Case Studies from Germany (C. Von Hirschhausen, T. Beckers, and K. Mitusch, eds). Edward Elgar Publishing. Knops, H.P.A. (2003). Weighing ways of keeping the energy balance. In Proceedings, 26th Annual IAEE Conference, Prague, 4–7 June. Künneke, R. and Fens, T. (2007). Ownership unbundling in electricity distribution: The case of The Netherlands. Energ. Pol., 35, 1920–30.

EBL

92

Competitive Electricity Markets

Kwoka, John (2006). Restructuring the US power sector: A review of recent studies. Report Prepared for the American Public Power Association, Northeastern University, November. http://www.appanet.org/ﬁles/PDFs/RestructuringStudyKwoka1.pdf Littlechild, S.C. (2003). Electricity: Regulatory Developments from Around the World. IEA/LBS Beesley lectures on regulation series XI, 9 October 2001, reprinted in Colin Robinson (ed.) Competition and Regulation in Utility Markets, London: Institute of Economic Affairs and London Business School, 2003, pp. 61–87. Meritet, Sophie (2007). French energy policy in the European context. For. Pol. in Dial., 8(20), 25–34. Newbery, D.M. (2002). Regulatory Challenges to European Electricity Liberalisation, Cambridge: DAE Working Papers. Newbery, D. (2005a). Electricity market reform in the European Union: Review of progress towards liberalization & integration. The Energ. J., European Energy Liberalization Special, 155–79. Newbery, D. (2005b). Reﬁning market design. Paper presented at the Conference on Implementing the Internal Market of Electricity: Proposals and Time-Tables, 9 September, Brussels, Sessa, http://www.sessa.eu.com/public/publications.php North, D. (1990). Institutions, Institutional Change And Economic Performance. Cambridge: Cambridge University Press. Parker, M.J. (2000). Thatcherism and the Fall of Coal. Oxford University Press. Potts Voll, S., Rosenzweig, M.B., and Pabon-Agudelo, C. (2006). Power sector reform: Is there a road forward. The Elec. J., 19(6), 24–37. Rosenzweig, M.B., Potts Voll, S., and Pabon-Agudelo, C. (2004). Power sector reform: Experiences from the road. The Elec. J., 17(11), 16–28. Rudnick, H., Barroso, L.A., Skerk, C., and Blanco, A. (2005). South American reform lessons: Twenty years of restructuring in Argentina, Brazil, and Chile. IEEE Pow. and Energ. Mag., July/August, 49–59. Scherer, F.M. (1980). Industrial Market Structure And Economic Performance, 2nd edn. Boston: Houghton Mifﬂin Comp. Serrallés, R.J. (2006). Electric energy restructuring in the European Union: Integration, subsidiarity and the challenge of harmonization. Energ. Pol., 34, 2542–51. Sioshansi, F.P. and Pfaffenberger, W. (eds.) (2006), Electricity Market Reform: An International Perspective, Oxford: Elsevier. Sioshansi, P. (forthcoming) Electricity markets and the ‘reform of the reforms’. Int. J. of Glob. Energ. Iss., x, No x. Shuttleworth, G. (2000). Opening European Electricity and Gas Markets. NERA. Stoft, S. (2002). Power System Economics, Designing Markets for Electricity. Piscataway, NJ: IEEE Press. Stiglitz, J. (1986). Economics of the Public Sector. New York: W.W. Norton and Company. Thomas, S. (2003). The seven brothers. Energ. Pol., 31(15), 393–403. Thomas, S. (2006). The grin of the Cheshire cat. Energ. Pol., 34(15), 1974–83. Tompson, W. (2004). Restructuring Russia’s electricity sector: towards effective competition of faux liberalisation?, Economics Department Working Papers No. 403 ECO/WKP(2004)26, OECD, Paris. Tongia, R. (2003). The Political Economy of Indian Power Sector Reforms. Working Paper #5. Program on Energy and Sustainable Development, Center for Environmental Science and Policy, Stanford Institute for International Studies, November. Van Damme, E. (2005). Liberalising the Dutch electricity market. The Energ. J., European Energy Liberalization Special, 155–79. Victor, D.G. (2006). Power sector reform in developing countries: Lessons from a comparative study of Brazil, China, India, Mexico and South Africa. World Forum on Energy Regulation III, Washington DC, 8–11 October. Victor, D.G. and Heller, T.C. (2006). The Political Economy of Power Sector Reform: The Experiences of Five Major Developing Countries. Cambridge: Cambridge University Press. Williams, J.H. and Ghanadan, R. (2006). Electricity reform in developing and transition countries: A reappraisal. Energ. Pol. 31, 815–44. Weinmann, J. (2007). Agglomerative magnets and informal regulatory networks – Electricity market design convergence in the USA and Continental Europe. Working Paper, Florence School of Regulation.

EBL

Hybrid Electricity Markets

93

Weinmann, J. and Bunn, D. (2004). Resource endowment and electricity sector reform. London Business School, Working Paper, Energy Markets Group. Williamson, O.E. (1998). Transaction cost economics how it works; where it is headed., De Economist, 146(1), 23–58. Willman, P., Coen, D., Currie, D., and Siner, M. (2003). The Evolution of regulatory relationships: regulatory institutions and ﬁrm behavior in privatized industries. Ind. and Corp. Ch., 12(1), 69–89. Yi-chong, X. (2006). The myth of the single solution: Electricity reforms and the World Bank. Energy, 31, 802–14.

EBL

EBL

This page intentionally left blank

Chapter 3 Achieving Electricity Market Integration in Europe NIGEL CORNWALL Managing Consultant, Cornwall Energy Associates

Summary This chapter describes the evolution toward the single electricity market in Europe under the auspices of the European Union (EU) following over a decade of reform. Its focus is on a number of centrally driven initiatives, but speciﬁc technical and regional initiatives to market integration are also examined. The chapter also looks at convergence of key market and network access activities, and considers collaboration between transmission system operators (TSOs) who initially took the lead. It seeks above all to synthesize a lot of sometimes confusing activity, disparate initiatives, and abundant literature into a single account. The chapter concludes that there are now reasonable prospects of achieving incremental and possibly sustained change. However, while there have been some successes, the reform path continues to prove much more difﬁcult than was initially envisaged. At this stage, while the likely direction of the evolutionary path toward functional and market harmonization is clear, it is not obvious when and in what precise form further convergence between markets will take.

EBL

3.1. Introduction European Union (EU) policymakers have been trying to develop an integrated European market. This chapter describes the evolution toward the single electricity market in Europe under the auspices of the EU following a decade of reform. The chapter’s focus is on a number of centrally driven initiatives, but speciﬁc technical and regional initiatives to market integration are also examined. It also looks at convergence of key market and network access activities and considers technical collaborations between transmission system operators (TSOs) who initially took the lead. Section 3.2. sets the contextual information on the European interconnected system and some of the key players. Section 3.3. runs through four distinct, though sometimes overlapping, milestones or phases of activity: • •

the ﬁrst phase concerned work by the TSOs in a number of regional sub-groupings; the debate on market convergence in Europe then picked up momentum under the stimulus of the ﬁrst electricity directive in 1996, with the European 95

Competitive Electricity Markets

96

Commission (hereafter the commission) taking over the driving seat for a second phase; • since the turn of the century the commission has become much more directional, and, with a second electricity directive in 2003 and regulations issued under it, it has reinvigorated a number of important workstreams, which effectively constitute a third phase; and • a fourth and new phase of market integration has recently begun. The commission has actively co-opted national regulators and refocused its efforts to help bring about sub-regional markets as a staging post to its one market vision. It has also initiated debate on further key measures as part of its January 2007 energy package. Under the stimulus of the commission and regional technical and regulatory agencies, the discussion by early 2007 was deﬁnitely back in full swing after a number of false dawns. However, the conviction that liberalization, market opening, price deregulation, and functional unbundling on their own would lead to competition and lower prices has long since dissipated in the commission and among those supporting the reform process. Section 3.4. takes a detailed look at key workstreams that have allowed some of the technical barriers to market convergence to be addressed. Some of these have been brought into clearer focus by the second directive and supporting regulation and guidelines. Section 3.5. examines how one sub-regional grouping comprising Britain, France, and Ireland (both north and south) is currently collaborating in a regional market initiative being led by the commission. There is relatively abundant literature available on these states, and the regional initiative involving them is making early progress. Unlike previous initiatives, this effort is being managed by the participating national regulators. The chapter’s main conclusions are summarized in Section 3.6.

EBL

3.2. Contextual Setting This section looks at the physical background, the key players, and the development of pan-European trade and mechanisms and is organized into ﬁve subsections: • • • • •

an overview of the European grid; a primer on the European TSOs; a discussion on trading convergence; an introduction to European exchanges; and a discussion of technical collaborations among the TSOs.

3.2.1. An overview of the European grid The electricity networks of the 25 member states at end-December 20061 are interconnected but to widely different degrees. A stylized representation is presented in Fig. 3.1.2 1

From 1 January, Romania and Bulgaria also accessed to the union. Original 15 member states: IE=Ireland, UK=United Kingdom, NL=Netherlands, DK_W=Denmark West, DK_E = Denmark East, NO = Norway, FI = Finland, SE = Sweden, DE = Germany, BE = Belgium, FR = France, ES = Spain, PT = Portugal, IT = Italy, AT = Austria, GR = Greece, Luxembourg not shown. Acceded prior to 2006: PL = Poland, EE = Estonia, LV = Latvia, LT = Lithuania, CZ = Czech Republic, SK = Slovakia, SL = Slovenia. HU = Hungary, Acceded 2007: RO = Romania, BA = Bulgaria. South East TSOs outside European Union: MK = Macedonia, AL = Albania. Others: Russia, BY = Belarussia, UA = Ukraine, TR = Turkey, MA = Morocco, and CH = Switzerland, are not part of the EU but are interconnected. 2

Achieving Electricity Market Integration in Europe

97

EBL

Fig. 3.1. Representation of European cross-border power ﬂows, 2005. Source: ETSO 2006.

Presently ﬂows between national markets are very limited. Within the EU cross-border trading of electricity is more important than exchange with countries outside the EU. Luxembourg, Latvia, and Hungary have net imports of 62%, 51%, and 22%, respectively, of their national consumption. At the other end of the spectrum sit the Czech Republic and Estonia that have net exports amounting to 31% and 41%, respectively, of their domestic consumption, whereas Lithuania’s net exports are, at 106%, higher than its domestic consumption. In terms of volumes the largest net exporter of electricity is France, which exported 67T Wh in 2003, four times more electricity than the next largest net exporter, the Czech Republic whose exports, however, grew 23-fold since 1990. Poland is third in this ranking. Italy was by far the most important net importer of electricity, importing approximately three times as much as the Netherlands, with Sweden coming as third largest net importer. Full data for 2004 and 2005 are shown in Table 3.1. Flows between markets are inhibited by wide-ranging differences in the access arrangements on different interconnectors or links. Even where a similar type of mechanism has been applied, timescales and details of allocation can differ widely.3 3 http://ec.europa.eu/energy/electricity/benchmarking/doc/2/sec_2003_448_en.pdf, see especially table page 81. This chapter provides detailed footnotes with relevant sources rather than a bibliography at the end.

Competitive Electricity Markets

98

Table 3.1. Electricity imports and exports, 2004 and 2005

Area/Country EU-25 EU-15 New -10 Austria Belgium Cyprus Czech Republic Germany Denmark Estonia Spain Finland France United Kingdom Greece Hungary Ireland Italy Lithuania Luxembourg Latvia Malta Netherlands Poland Portugal Sweden Slovenia Slovakia

Total electricity demand

Imports from foreign countries

TWh

MW

Exports to foreign countries TWh

2005

2004

2005

2004

2005

3 0135 2 7108 3926

2893 2458 435

3015 2509 506

2942 2235 707

2811 2072 739

638 882 41 250 5635 352 74 2716 849 4824 3866 584 388 265 3294 101 62 68 23 1147 1306 521 1473 137 262

166 146 00 −09 442 87 03 01 117 291 98 49 138 15 465 01 65 56 00 214 53 65 156 43 43

204 143 00 98 534 134 03 101 179 00 113 56 150 21 502 11 64 51 00 237 50 75 146 72 45

135 60 00 124 215 115 23 111 68 910 23 20 63 00 08 73 31 35 00 52 146 01 178 50 62

177 80 00 255 619 121 21 11050 09 602 00 18 88 00 141 41 32 29 00 54 162 07 220 76 72

EBL

Source: Eurelectric, March 2007.

Most existing links between member states are also heavily constrained. Congestion management methods also differ widely, with a number of variations apparent and typically based around three different models, namely re-dispatch, counter-trading, and multilateral transmission capacity allocation. A key objective of the single pan-European electricity market is to increase the extent of interconnection and levels of trade between markets, and the commission has long recognized that interconnectors facilitate the inter-regional and cross-border transport of power and that a high level of interconnection is a pre-requisite for a well-functioning internal market. It was agreed at the Barcelona European Council in 2002, following a range of ﬁtful and largely unsuccessful initiatives over the 1990s, to strengthen policy to facilitate the completion of infrastructure projects seen by member states as priorities.

Achieving Electricity Market Integration in Europe

99

The need for a strengthened policy to facilitate the completion of priority infrastructure projects was underlined by the EU heads of state and government at Hampton Court in October 2005. Ministers agreed to formalize a policy to increase minimum interconnection levels to 10%. However, a signiﬁcant number of member states have not achieved this target.4 Consequently, the EU has formulated a series of policies aimed at supporting the development of effective energy infrastructure in Europe. Those impacting on electricity include: Guidelines for trans-European energy networks (the so-called TEN-E Guidelines)5 – the commission has identiﬁed 314 projects of common interest whose completion, it considers, should be speeded up. These include 42 high-priority “projects of European interest,” which may be either cross-border in nature or have a signiﬁcant impact on cross-border transmission capacity. The Guidelines provide a framework for increased coordination, for monitoring progress in implementation, and where appropriate, for ﬁnancial support, usually in the form of loans by the European Investment Bank. • Speciﬁc rules to ensure an appropriate level of electricity interconnection between member states, while facilitating a stable investment climate as mandated by a 2005 directive.6 •

Despite these and other measures, the commission continues to consider that progress on development of common networks is insufﬁcient. The European Council of March 2006 called for the adoption of a Priority Interconnection Plan7 (the Plan), as part of the Strategic European Energy Review (SEER). And this was issued in January 2007. This plan addresses how the priority projects can be stimulated, especially the 42 high-priority schemes, 32 of which are electricity schemes out of which 20 have faced signiﬁcant delays. In all the commission considers that some E6 bn of investment is necessary by 2013. The issues have taken on more immediacy following the blackout that occurred in eight EU

EBL

4

These include Poland, the United Kingdom, Spain, Ireland, Italy, France, Portugal as well as Bulgaria and Poland. 5 Decision No 1364/2006/EC (OJ L 262, 22 September 2006, p.1). 6 Directive to safeguard security of electricity supply and infrastructure investment, 2005/89/EC. 7 The Priority Interconnection Plan sets out ﬁve priorities: • identifying the most signiﬁcant missing infrastructure up to 2013 and ensuring pan-European political support to ﬁll the gaps; • appointing four European coordinators to pursue the four of the most important priority projects, three on which are in the power sector: the Power-Link between Germany, Poland, and Lithuania; connections to offshore wind power in Northern Europe; electricity interconnections between France and Spain; • agreeing for a maximum of 5 years within which planning and approval procedures must be completed for projects that are deﬁned as being “of European interest” under the commission’s TEN-E Guidelines; • examining the need to increase funding for the Energy Trans-European networks, particularly to facilitate the integration of renewable electricity into the grid; and • establishing a new Community mechanism and structure for TSO, responsible for coordinated network planning. http://eur-lex.europa.eu/LexUriServ/site/en/com/2006/com2006_0846en01.pdf

100

Competitive Electricity Markets

countries on 4 November 2006 affecting 15 mn customers,8 which highlighted that the European grid is already behaving in some important respects as a single power system, but with the network not designed accordingly.

3.2.2. The European TSOs In Europe, as in other large interconnected systems, TSOs have been key players in stimulating market convergence. A TSO is a Transmission System Operator, in principle, similar to regional transmission organizations (RTOs) in US parlance. Its job in a nutshell is to “keep the lights on,” within prescribed statutory and technical limits over the part of the network it has jurisdiction over. The European Transmission System Operator (ETSO) organization, which loosely acts as a trade association for them, deﬁnes a TSO as the party who is: responsible for the bulk transmission of electric power on the main high voltage electric networks. TSOs provide grid access to the electricity market players according to non-discriminatory and transparent rules. In order to ensure the security of supply, they also guarantee the safe operation and maintenance of the system. In many countries, TSOs are in charge of the development of the grid infrastructure too. TSOs in the European Union internal electricity market are entities operating independently from the other electricity market players.9 Typically, each country in Europe has its own national TSO, though some have more than one (e.g., Germany), usually in areas where regional utilities have historically been the industry model. Most TSOs own the transmission system, some don’t, but they all exercise operational control over high-voltage networks, though actual boundaries and demarcations differ.10 In many jurisdictions – and all of those this chapter is most interested in – the TSO is also responsible for trading and settlement of energy imbalances where parties inject or withdraw power from the physical system without energy contracts. While there is great diversity in form and to a lesser degree function, the degree of unbundling achieved is lesser than that evident in most parts of North America and Australia. There also tends

EBL

8 In its advice delivered at the request of the commission on 20 December 2006, ERGEG concluded that lessons from an earlier 2003 Italian blackout have not been followed through, and that the following was needed to keep the lights on in Europe in the future: • adoption, on proposal of the European Commission, of legally binding operational security rules; • development by the commission of a framework for the electricity network as part of its energy strategy; • improvement of the cooperation between EU electricity grid operators, which should be publicly accountable for their actions. http://www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_ADMIN/PR-07-03_ERGEG_ FinalReport_Blackout_2007-02-06.doc Ergeg’s Final Report identiﬁes three main causes: (1) non-fulﬁlment of the (n-1) security rule, (2) inappropriate inter-TSO coordination during the event and (3) distributed generation units were not monitored or controlled appropriately by TSOs The uncoordinated behavior worsened the effects. 9 www.etso-net.org/association/aboutus/tso/e_default.asp. 10 In all member states, voltage levels of 220 kV and above are included as transmission. There exists a range from 50 kV to 200 kV where the voltage level is ether transmission or distribution. All voltage levels below 50 kV are included in the distribution networks.

Achieving Electricity Market Integration in Europe

101

to be less differentiation between the respective roles of asset owner and system operator than in other deregulated electricity markets. TSOs undertake a number of essential functions, which are typically performed under legislative or regulatory provisions. The precise way in which TSOs achieve these functions varies, but the core activities are: •

• • •

•

•

network capacity and investment – TSOs manage network investment to ensure that the infrastructure continues to meet technical security standards and to ensure that sufﬁcient capacity is delivered; network access – TSOs contract with grid users to provide access to their respective transmission networks; transmission charging – TSOs also levy charges on grid users to recover the costs of the network; network operation – TSOs operate the transmission system in real time and carry out congestion management and procure balancing or ancillary services to help them do this; emergency planning and, in extremis, system restart – TSOs plan and make arrangements for possible network disturbance situations such as the failure of a major generating station or the more widespread failure of part or all of the network. Each TSO will establish operational plans to recover from these situations and will enter into contracts for the procurement of adequate services needed for recovery, e.g., black start of generators, reserves etc; and network maintenance – TSOs undertake the day-to-day maintenance of their transmission network equipment.

EBL

Managing access to the system is also a key activity. This entails providing a contractual interface with grid users, and developing rules for connection to the system and its use. As well as applying terms for grid use, TSOs levy charges not just for use of the assets but also to deal with congestion and thermal network losses. Table 3.2 lists the key TSOs in Europe. The markets most relevant to this chapter are in the top four country rows. Electricity systems across the countries of Europe have many shared characteristics, and the measures TSOs adopt to maintain balance of the electricity system also have common drivers. The operation of TSOs and the technical parameters that apply in Europe is well documented,11 12 13 14 and is not covered further in this chapter. 3.2.3. Deregulation The uniformity of the timetable for deregulation conceals important differences at which the pace of change has been introduced by different member states. The size and composition of national markets also differ widely. Basic data are shown in Fig. 3.2. The red part of the bar shows the extent of the market for non-domestic customers’, dark blue the part of the domestic retail market that has been opened, and light blue the part of 11

ETSO (2003). Current State of Balance Management in Europe. December. Frontier Economics Ltd for the European Commission (2005). Beneﬁts and Practical Steps Towards the Integration of Intraday Electricity Markets and Balancing Mechanisms. December. 13 Ergeg (2006). Guidelines of Good Practice for Electricity Balancing Markets Integration. 7 June. 14 Verhaegen, K., Meeus, L., and Belmans, R. (2006). Development of balancing in the internal electricity market in Europe.presented at EWEC 2006 Athens, 28 February. 12

Competitive Electricity Markets

102

Table 3.2. European TSOs Country

Transmission system operator

Grid operator/ owner

Imbalance settlement

Power exchange

Britain

NGET

NGET (England and Wales) SSE (North Scotland) SP (South Scotland)

Elexon

APX (UK)

Ireland

EirGrid

ESB NG

EirGrid

Northern Ireland

SONI

France

RTE

Powernext

Denmark

Energinet.dk

NordPool

Finland

Fingrid

Norway

Statnett

Sweden

Svenska Kraftnat

Austria (APG)

VERBUND APG

Austria (Tirol)

TIRAG

Austria (Vorarlberg)

VKW Netz

Belgium

Elia

Germany

EnBW Transportnetze AG, E.ON Netz GmbH, RWE Transportnetz Strom GmbHNET, Vattenfall Europe Transmission GmbH

Greece

HTSO

HTSO / PPC

HTSO

Hungary

MAVIR Rt. (ISO)

MVM Rt.

MAVIR Rt.

Italy

GRTN

13 owners with TERNA >95%

GME/GRTN

Luxembourg

Cegedel

Netherlands

TenneT

Poland

Polskie Sieci Elektroenergetyczne SA

Portugal

REN

EBL

APCS

EXAA

A&B

EEX

Towarowa Gielda Energii SA

Achieving Electricity Market Integration in Europe

103

Table 3.2. (Continued) Country

Transmission system operator

Grid operator/ owner

Imbalance settlement

Power exchange

Romania

C.N. Transelectrica S.A.

Slovak Republic

Slovenska elektrizacna prenosova sustava, a.s.

Slovenia

ELES

BORZEN (controlled subsidary of ELES)

BORZEN

Spain

Red Electrica de Espana (REE)

Compania Operadora del Mercado de Electricidad (OMEL)

Source: Cornwall Energy, 2007. DE

556.4

FR

482.4

UK

355.0

IT

329.1

ES

Total national market volume(TWh)

247.9

SE

147.3

PL

130.6

NO

EBL

125.9

NL

114.7

BE

86.8

FI

85.0

AT

63.2

CH

63.0

CZ

62.7

GR

52.9

PT

49.9

HU

Opened I&C market

39.3

DK

35.7

SK

26.3

IRL

23.7

SI

Non eligible I&C market Opened retail market Non eligible retail market

12.8

LU

6.2 0

100

200

300 Annual volume (TWh)

400

500

600

Fig. 3.2. Market size and contestability. Source: EU data as consolidated by Cap Gemini.

the domestic market that remains closed prior to 1 July 2007 when full market opening is set to occur. It shows that in 10 of the original group of 15, full competition has been established legally. Actual competition is a rather different matter and is not the focus of this chapter. 3.2.4. Trading convergence Technical collaboration and convergence has been accompanied by continuing divergence in regulatory and commercial rules across national markets. Trading between participants in the different European electricity markets also takes place over different timescales,

Competitive Electricity Markets

104

though a common broad structure is emerging in most jurisdictions now. This standardization has developed in the main without the unifying pressures seen across North America and Australia through federal initiatives and orders covered in other chapters of this volume. Of recent power exchanges (PXs) have emerged and, through the trade association Europex, a pan-European trade association that now has 14 members representing existing or planned national exchanges, they have begin to express a common voice on trading matters and convergences between markets. EFET, the European Federation of Energy Traders, has also acted as a focus for energy traders. In both cases the trade associations have produced their own agendas and priority lists for action to help expedite the single market not just for electricity but also for gas. The basic components of the electricity trading model that is emerging is summarized in Fig. 3.3. It is explained in more detail in Box 3.1.

Year (s) Ahead

Month/Week Ahead

Day Ahead

In Day

Settlement

Balancing Period

Point of Delivery

Gate Closure Bilateral Over the Counter (O.T.C.) Trading: Hedging Trades Against Generation/Supply Speculative Large Granularity (Volume and Time Range)

EBL

Anonymous Power Exchange (P.X.) Trading: Hedging/Refining Position Balancing Trades Medium Granularity (Volume and Time Range)

P.X. / Bilateral Trading: Refining Position Balancing Trades Small Granularity

System Operator (T.S.O.) Trading: System Operator strikes long term contracts for the provision of system services (to ensure capacity to provide services is available). The method of procurement may extend to organized – complementary – markets for the procurement of specific products

Balancing mechanisms and balancing markets T.S.O. buys balancing energy and system services and calls off services under longer term contracts

Fig. 3.3. Typical timescales for trading of energy. Source: Based on a diagram from Beneﬁts and Practical Steps Towards the Integration of Intraday Electricity Markets and Balancing Mechanisms, December 2005, modiﬁed by Cornwall Energy, March 2007.

Achieving Electricity Market Integration in Europe

105

Box 3.1 Key elements of market design – European style There are six core elements to the basic trading model. First, there are bilateral and over-the-counter (OTC) markets, where the bulk of energy wholesale trades occur (typically up to over 95%). Trading usually occurs prior to the day-ahead stage, and it often stretches a year or more ahead. With an OTC trade, the transaction is facilitated by a third-party agent. Second, there are the organized exchanges, which are anonymous. They are usually optional (APX in Britain, EEX in Germany, Powernext in France) or semi-optional in the sense that trading is obligatory for un-contracted quantities (NordPool, APX in Netherlands), where standardized contracts ranging from the very short term (day-ahead) to the medium term (1-month to 2-year futures) are traded. Within the EU, these organized regional markets account for between 1 and 2% of the energy consumed as in the United Kingdom to 20% (maximum) as in NordPool. Third, there are different forms of TSO trading to manage network constraints and ensure system balance. These include congestion management mechanisms, which may function as energy markets for allocating transmission capacity (as in the NordPool, which breaks into different price zones when a transmission constraint occurs), or as an organized, single-buyer market (the TSO purchasing on the balancing “countertrading” market), or in the form of a bilateral contracting with the TSO negotiating a portfolio of contracts (re-dispatching). In Britain a hybrid form of counter-trading and re-dispatch is adopted. They also include balancing mechanisms, which give TSOs the means for real-time balancing of energy injections and withdrawals from the physical system. These mechanisms may rely on bilateral contacts negotiated by the TSO or on an organized market for the purchase and sale of incremental and decremental energy. Where energy imbalances are discouraged by applying a penalty on the cost of energy balancing, the term “balancing mechanism” is often used. If, on the other hand, energy is sold (upward or downward as in France) at cost and if parties are permitted to purchase (upward or downward), the term “balancing or adjustment market” is usually used. Further, there are complementary markets, such as ancillary service markets and capacity markets, under which a TSO may provide centralized markets with the means to ensure their functioning or contribute to both short- and long-term security of supply. Generally speaking, these complementary markets in Europe are not well developed compared with North America in the sense that they tend not to be open, with TSOs instead having wide discretionary powers to procure reserves and balancing energy bilaterally. Spain is a notable exception. Finally, there are retail markets. As summarized in Section 3.2.3. these are showing common and converging characteristics. All markets within the EU had to introduce competition to all non-domestic customers by July 2004. Full retail competition is to be introduced by July 2007. Despite efforts by the commission to benchmark competitive behavior within member states, retail markets are still governed by a signiﬁcant diversity of rules, and they are a considerable way from being subject to a uniform regulatory retail framework.

EBL

Competitive Electricity Markets

106 3.2.5. European exchanges

Trading on organized exchanges in Europe has been slow to take off outside of the Scandinavian market NordPool primarily because of the high levels of vertical integration and consolidation that prevail but, in its preliminary report from its sector investigation, the commission was upbeat about recent improvements and the prospects for increased liquidity in traded markets. Figure 3.4 shows the development of traded spot volumes relative to the consumption in the relevant geographical area for selected markets. Table 3.3 shows spot volumes traded on PXs and on OTC markets relative to electricity consumption in the relevant geographical area. It is evident that large differences exist between geographical areas. These differences are partly the result of diverging national wholesale market frameworks. PXs, as previously noted, can be divided into two broad groups. In the ﬁrst group, members of PXs have some requirement or incentive to trade via the exchange (OMEL, GME, NordPool), and which as a consequence show a high percentage of trading. In the second group exchange members have no such incentives and open trading in electricity is in effect only just emerging. In this group EEX and APX saw signiﬁcantly higher spot volumes traded than Powernext, EXAA, Pol PX, and the UKPX. Because of these differences a direct comparison between the two groups of exchanges reveals little, but from Table 3.3 it also emerges that traded spot volumes on exchanges are larger than brokered spot markets in most of the countries. As can be seen from Table 3.4, total traded volumes in standardized forward contracts show large variations among countries, suggesting varying degrees of market development. Again, market design is an important factor. Forward trading in Spain is insigniﬁcant, reﬂecting the mandatory nature of the pool system. In contrast, the Dutch and German OTC forward markets traded by far the highest volumes (relative to consumption) on the continent. In terms of trades a number of continental markets saw their volumes rise. The German and the Dutch markets in particular experienced increasing OTC volumes.

EBL

Spot volumes are developing Development of spot traded volumes on selected power exchanges as a percentage of national consumption

18% 16% 14% 12% 10% 8% 6% 4% 2% 0% 2000

2001 Powernext

2002

2003 APX

2004

2005 EEX

Fig. 3.4. Traded volumes on spot markets. Source: European Commission (2006). Preliminary sector report. February.

Achieving Electricity Market Integration in Europe

107

Table 3.3. Spot traded volumes as a percentage of national electricity consumption (June 2004– May 2005)

Powernext – France UKPX – UK OMEL –Spain GME – Italy NordPool – Nordic Region EEX – Germany APX – The Netherlands Belgium EXAA – Austria Pol PX – Poland

Power exchanges

OTC brokered

3.37% 2.17% 84.02% 43.67% 42.82% 13.24% 11.88% No power exchange 2.96% 1.28%

1.50% 8.60% Negligible n/a n/a 5.40% 5.90% 0.04% n.a n.a

Source: European Commission (2006). Preliminary sector report. February. Note: This table does not contain an exhaustive list of all power exchanges in Europe. OTC brokered numbers refer to volumes reported to the commission by major energy brokers. Table 3.4. Traded volumes in futures/forward contracts as a percentage of national electricity consumption (June 2004–May 2005) Power exchanges

OTC brokered

OMEL – Spain

No exchange trading

Negligible

n/a

GME – Italy

No exchange trading

n/a

n/a

NordPool – Nordic region(2004)

151%

n/a

n/a

EEX – Germany

74%

565%

639%

Endex – The Netherlands (since December 2004)

39%

509%

548%

Belgium

No exchange trading

22%

22%

Powernext – France

6%

79%

85%

EXAA – Austria

No exchange trading

n/a

n/a

Pol PX – Poland

No exchange trading

n/a

n/a

UKPX – UK

0%

146%

146%

EBL

Power exchange + OTC

Source: European Commission (2006). Preliminary sector report. February.

While the development of total traded volumes as a proportion of national electricity consumption has increased, the absolute levels of trading need to be kept in perspective. The United Kingdom is the only market in the sample where traded volumes have signiﬁcantly declined during the last 2 years. This development is ascribed by the commission to vertical reintegration of the industry, i.e., the trend to combine independent generation and supply businesses into a single operation under the same ownership. Volumes continue to be quite low in France and in Belgium owing to the high level of concentration and vertical integration in these countries. European power trading experienced its best year ever in 2005. In 2005, European PX volumes were up 30%, reaching a record high of 3781 TWh, representing 126% of total European power consumption (Box 3.2). The increasing number of exchanges,

108

Competitive Electricity Markets

standardization and simpliﬁcation of rules, and ofﬁcial encouragement all played their part. Spot volumes in major West European PXs reached 734 TWh, increasing by 38% over 2004. European spot volumes also increased by 20% between Winter 2005–06 and Winter 2004–05, which may point to a conﬁrmation of the upward trend in 2006.

Box 3.2 European power exchanges NordPool NordPool was established in 1991. It has progressively expanded from Norway, to Sweden then Finland and Denmark. It remains by far Europe’s largest PX in terms of volumes. In 2005, its total spot, futures, and OTC volumes rose by 27% to 962 TWh, 50% of West European PX trading in 2005. EEX Germany’s EEX is now Europe’s second largest PX. It recorded total spot and futures contract volumes of 347 TWh in 2005, up 52% on the year. EEX launched in summer 2005 an OTC clearing offer for French power forwards, allowing the two markets to be connected, the ﬁrst joint clearing procedure in Europe. Omel

EBL

Omel is the pool operator in Spain. The exchange, established in 1997, is mandatory. Its day-ahead volumes rose by 11% to 223 TWh in 2005, making Omel Europe’s largest spot market. IPEX In its second year of trading (the exchange opened in 2004), Italy’s IPEX’s day-ahead market recorded volumes of 203 TWh, making it Europe’s largest spot market after Spain’s Omel. Again, trading through it is not discretionary for most participants in the Italian market. Powernext Powernext commenced operations in 2002. Spot market volumes on France’s Powernext jumped by 39% in 2005 to 19.7 TWh, whereas futures volumes increased by a factor of three from 13 TWh in 2004 to 62.4 TWh in 2005. Endex Endex offers Dutch power futures and clearing of Dutch OTC power forward contracts and Belgian power forwards. In 2005 Endex booked total volumes of 106 TWh, up 150% on the previous year. Most of the trading activity was Dutch. The creation of a regional French–Belgian–Dutch day-ahead power market and single trading zone, Belpex, was also approved by the Belgian Government in January 2006. Belpex will establish a single trading zone. The project will also create

Achieving Electricity Market Integration in Europe

109

a single power price for France, the Netherlands, and Belgium in the absence of constraints. East European Power exchanges There are four other operating PXs in Eastern Europe. They serve the Czech Republic, Poland, Romania, and Slovenia. All the exchanges presently offer spot market services alone. In 2005, the combined volumes of the Eastern exchanges were 4 TWh, accounting for 2% of total electricity consumption in the four countries.

Integrated regional market areas are developing in spite of remaining regulatory and political uncertainties. In 2005, different market prices were highly correlated15 despite the fact that the implementation of cross-border regulation was lagging, as explained in Section 3.5. However, regional trading structures are very varied and, outside of NordPool, the development remains a signiﬁcant way from the development of strong regional trading hubs. 3.2.6. Technical collaborations TSOs, particularly those that are part of a highly meshed and synchronized network, cooperate with other TSOs to ensure the secure technical operation of their joint networks. This is not just a European phenomenon, as the evolution of security coordinating councils in North America illustrates. In Europe signiﬁcant progress has already been made under the aegis of pan-European technical agencies toward a more harmonized grid over a period in excess of ﬁfty years. There are now four groupings in Central and Western Europe where TSOs cooperate together or where they have formed formal collaborations. These areas are:

EBL

UCTE,16 which is a grouping of mainland European and some northern African countries; • Nordel,17 which is a grouping of the four Scandinavian countries; •

15

German and French prices were 99.69% correlated, up from 98.28% in 2004, and German and UK 96.27%, up from 79.75%. 16 The Union for the Co-ordination of Transmission of Electricity was originally set up in 1951 with seven members, Austria, Belgium, France, Federal Republic of Germany, Italy, Luxembourg, and the Netherlands followed by the connection of Denmark. In 1987 Portugal, Spain, Yugoslavia, Greece, and Albania were connected to the system, although the southeastern European states were disconnected during the Balkan wars of the 1990s. In 1995 the CENTREL system countries of Poland, Czech Republic, Slovak Republic, and Hungary were synchronized. In 1996, Romania and Bulgaria were connected. In 1997 the Maghreb countries Morocco, Algeria, and Tunisia were connected via the Gibraltar interconnector, and the Western Ukraine was connected in 2003. Currently UCTE is investigating the addition of Turkey and the eastern Mediterranean to the system, a project it calls “completing the ring.” 17 Nordel is a body for cooperation between the TSOs in Denmark, Finland, Iceland, Norway, and Sweden, whose primary objective is to create the conditions for, and to develop further, an efﬁcient and harmonized Nordic electricity market. Iceland is also a member of Nordel but there is no interconnection to the other countries, the closest of which is Norway at 970 km. Denmark also has a foot in the UCTE camp and the western part of the country is therefore synchronized with UCTE.

110 • •

Competitive Electricity Markets Island of Ireland, which is Northern Ireland and the Irish Republic;18 and Island of Britain, which is the English and Welsh market and the Scottish market.19

With the emergence of the commission’s policy to create a single electricity market, the TSOs of the four regional market groupings decided to set up the international association ETSO to accelerate the harmonization of network access and conditions of usage to facilitate cross-border electricity trade. Added to these regional technical initiatives, ETSO has looked across these regional groupings and provided a forum for cohesion and achieving consistency on a wide range of technical and access issues, from the deﬁnition and speciﬁcation of system services to information shared protocols and common disclosure. More recently ETSO has also progressed several areas of work under wider programs, including inter-TSO compensation (ITC) to deal with cross-border ﬂows, congestion management, electronic data interchange, network use tariffs, security of supply, integration of renewables, and balance management, which is described in more detail later. It is also working closely with national regulators and the commission through the Florence Forum process established by the commission in 1998 to stimulate progress under the ﬁrst electricity directive. That said, the degree of convergence and consistency achieved between the various TSO entities is often overstated, and in a 2005 report the commission was able to highlight wide-ranging differences in the legal status and in the degree of unbundling achieved between them.20 Further, in February 2006 consultants retained by the commission carried out an inventory and comparative analysis of the transmission network security and reliability rules in Europe. The report concluded that there was a signiﬁcant diversity in the style and content of the codes and even “in the direct purposes for which they are written.” They recommended a general form of standardization of grid codes and a clearer legal hierarchy, and advocated a consistent scope and format, coverage, and governance arrangements using common terminology.21

EBL

3.3. European Commission Initiatives The commission is playing an increasingly important role in the development of closer working relationships between TSOs, regulators, and other market participants to achieve 18

The Island of Ireland collaboration is a relatively recent exercise. ESB was until recently responsible for all aspects of generation, supply, and infrastructure in the Republic of Ireland. In 2000 the market began to be opened up. On 1 July, 2006, a ring-fenced subsidiary EirGrid plc took over the TSO responsibilities from ESB, which still owns the infrastructure. SONI, the System Operator for Northern Ireland, is a wholly owned subsidiary of Northern Ireland Electricity and is the TSO for Northern Ireland. Discussions are progressing between EirGrid, SONI, and various regulators on the creation of an “All Island Electricity Market.” The new TSO for the whole island market will be called AIME. 19 Until 2005, the transmission systems in England and Wales and Scotland operated separately though there was a high-level technical protocol between the system operators on either side of the Scottish border. From April 2005 a single trading and transmission arrangement across Britain was implemented, termed Betta, which essentially reﬂected the rollout of the England and Wales Neta style market and a single GB system operator from England and Wales into Scotland. As part of these changes, a number of key changes were introduced that impacted on system operation, including uniform transmission access and pricing rules and appointment of a single TSO independent of market players, with the two Scottish utilities divesting themselves of system operation functions. 20 http://ec.europa.eu/energy/electricity/report_2005/doc/2005_report_technical_annex.pdf 21 http://ec.europa.eu/energy/electricity/publications/doc/security_rules_pb_power_february_ 2006.pdf.

Achieving Electricity Market Integration in Europe

111

a competitive, integrated internal market and in tackling barriers to harmonization. It has to, as its legal mandate is to establish a single energy market, which presupposes the free movement and trade of electricity within and between member states. Setting aside legalities, it has in recent years seized the initiative following haphazard progress by many member states in transcribing two landmark electricity directives into national law, and it is in the process of taking infringement proceedings against several member states.22 A series of benchmarking reports that began in 2001 on progress in implementation of the internal markets and addressing the practical results of the commission’s policies have also highlighted mounting dissatisfaction on the part of the commission at what it sees as some member states acting against the spirit as well as the letter of community law. “In summary, stakeholders do not have a high degree of conﬁdence in the internal market,” the commission wrote in January 2007,23 and it has under way a wide-ranging inquiry of the sector (with gas) under competition law. This section addresses three areas of commission activity that is designed to stimulate progress to the evolution of the single electricity market: •

the legislative background to electricity market restructuring established by the commission; and • other initiatives by the commission to foster harmonization, including establishment of a devolved regional process. 3.3.1. Legislation As Europe as a whole has moved to a closer political union, the electricity market has followed the trend though not as fast as the commission would like. A pivotal point in the development of a common policy was the enactment of the ﬁrst electricity directive in 1996.24 The ﬁrst directive provided for a gradual minimum market opening in three steps:

EBL

• • •

22

ﬁrst step on 19 February 1999; second step on 19 February 2000; and third step on 19 February 2003.

The commission took action against 20 member states for violation and non-transposition of the existing Directives. Following letters of formal notice sent in April 2006 and in advance, if needed, of starting procedure before the European Court of Justice, on 12 December 2006, the commission decided to send 26 reasoned opinions to 16 member states including all the biggest. The main deﬁciencies observed in transposition of the new internal market directives are the following: • regulated prices preventing entry of new market players • insufﬁcient unbundling of transmission and distribution system operators, which cannot guarantee their independence • discriminatory third-party access to the network, in particular as regards preferential access being granted to incumbents for historical long-term contracts • insufﬁcient competencies of the regulators • no information given to the commission on public service obligations, especially as regards regulated supply tariffs • insufﬁcient indication of the origin of electricity, which is essential in particular for the promotion of renewable energy 23 Communication from the commission to the council and the European Parliament on the prospects for the gas and electricity market, COM(2006) 841 ﬁnal, p. 2. 24 Directive 96/92/EC concerning common rules of the internal market in electricity was adopted by the Council of Ministers on 19 December 1996. It entered into force 2 months later on 19 February 1997.

112

Competitive Electricity Markets

These requirements, which essentially imposed reducing size bands for “eligible” customers (those that were able to choose supplier), opened up the non-domestic retail markets. The commission also concentrated on establishing minimal requirements for unbundling of generation and transmission, and contemplated standardization of approaches to network access and its regulation.25 This emphasis on providing access to the market dominated over issues of restructuring generation and supply and market design. The commission’s leadership of the implementation process has primarily been through the Florence Forum, which ﬁrst met in February 1998 and typically meets annually (curiously now in Rome), and its various sub-groups. The 13th full meeting took place in September 2006.26 Over recent years the forum has provided a focus for direction of ETSO’s deliberations and liaison with national regulators, and it has also stimulated the formation of the pan-European energy regulators’ group, the Council of European Energy Regulators (CEER). Limited progress toward the objective of an internal market without barriers has already been achieved. The deregulation timetable – that for permitting competition for customers – has tended to be the political focus, but national diversity has continued to be a strong characteristic. As one commentator has noted the ﬁrst directive allowed nearly everything except an integrated internal market.27 However, there have been other important initiatives, which have increasingly been absorbed under the supervision of the forum. For example, from 2002 ETSO introduced a mechanism for cross-border tariffs (CBTs) for transactions between participating member states, and which has now removed speciﬁc transmission charges associated with exchanging electricity across most of the internal borders of the EU. In addition, voluntary guidelines for congestion management were agreed at the sixth meeting of the Florence Forum, and these have been partially implemented. Finally, considerable technical work has been carried out in preparation for a more comprehensive integration of markets by ETSO, EuroPEX, EFET, and the UCTE, as well as by the CEER. Many of their detailed initiatives are now reaching more detailed expression through the work program of Ergeg, which is explored below.

EBL

Table 3.5. EU deregulation milestones, 1996–2007 1996

European Council of Energy Ministers and Parliament reached agreement on an electricity market liberalization directive

February 1997

This “Directive concerning common rules for the internal market in electricity” (Directive 96/92EC) took legal effect

February 1999

Directive went into force after a two-year transposition delay

2001

Approval of the “Directive of the European Parliament and the Council on the promotion of electricity from renewable energy sources in the internal electricity market (RES-E directive)”

2003

Approval of the Directive concerning common rules for the internal market in electricity (ofﬁcially directive 2003/54; usually named “the second directive”)

25

This comprised regulated or negotiated third-party access, but also permitted single buyer models. Florence Forum papers can be found at: http://ec.europa.eu/energy/electricity/ﬂorence/index_ en.htm. 27 Lee Hancher, quoted by J-M Glachant (2005). Implementing the European internal energy market in 2005–09, available at www.grjm.net. 26

Achieving Electricity Market Integration in Europe

113

Table 3.5. (Continued) 2004

Extension of the EU 15 to 25 member countries; new member countries to open their market with 30% minimum

2004

Electricity directive 2003/54 due to be transposed by member states All non-domestic customers made eligible in the EU in July 2004 An EU regulation on cross-border electricity trade came into effect (Regulation 1228/2003) in July 2004

2007

Due to Electricity directive 2003/54, 100% market opening in all EU-25 countries in July 2007

The gradual market opening introduced by the ﬁrst directive resulted in signiﬁcant differences between member states regarding the level of market opening. By 2003 some alignment had occurred. However, the non-eligible segment of the retail markets remained – and still remains – signiﬁcant in many markets. The existence of negotiated third-party access regimes, the limited level of functional unbundling obligations, and the lack of any requirement to establish a national energy regulator were also viewed as obstacles. To address these concerns further measures were proposed by the commission leading to the adoption of a second directive. The second electricity directive and an important accompanying regulation on crossborder exchanges were adopted by the Council and Parliament on 16 June 2003.28 Its primary focus was the removal of discrepancies in the level of market opening between member states. Among the measures required by it are market opening (i.e., full retail competition) by July 2007, legal unbundling of functions and the introduction of sectorspeciﬁc regulation, and the establishment of national regulators independent from the sectors they regulate in all member states in order to ensure non-discriminatory access to networks through dealing with complaints and controlling network tariffs. National regulators retain a key role in setting up and enforcing most of the aspects of market design, in particular by removing inappropriate local technical and ﬁnancial impediments. Similarly, legal and functionally independent TSOs will, by providing non-discriminatory access to networks, be responsible for the day-to-day functioning of the electric system. The directive also contemplated speciﬁc areas of focus in developing local blueprints to improve market access and functioning. At the same time the existing national regulators recognized the need for close cooperation especially with regard to cross-border trade, and formed CEER as an association for discussion and the development of common positions. The measures by which the success of the second directive is to be judged are set out in its Article 28, paragraph 3. They are:

EBL

• • •

the existence of non-discriminatory network access; effective regulation; the development of interconnection infrastructure and the security of supply situation; • the extent to which markets are open to effective competition; 28

The Electricity directive 2003/54/EC is the key European legislation to establish the internal market of electricity. The directive had to be implemented by the member states by 1 July 2004.

114

Competitive Electricity Markets

•

the extent to which the full beneﬁts of market opening are accruing to small enterprises and households, “notably with respect to public service and universal service standards”; • the extent to which customers are actually switching suppliers; • price developments; and • independence of TSOs. As such the criteria apply a framework that mixes political, regulatory, operation, and trading considerations. Subsequently, the regulation on cross-border electricity exchanges29 was introduced to implement aspects of the second directive, and it aims to foster cross-border trade by allowing the commission, working within a “comitology”30 procedure, to adopt three binding guidelines: the allocation of cross-border electricity interconnector capacity (i.e., congestion management), harmonization of transmission tariffs, and compensation for TSOs for hosting transmission ﬂows originating in other TSOs’ areas. Accordingly, the regulation allows for the adoption of speciﬁc binding guidelines for cross-border transactions. It allows the development of harmonized conditions of access to the European network for those wishing to buy or sell (or trade) electricity. It is intended to lead to cost-reﬂective charges for the use of European transmission networks, the removal of other distortions of cross-border trade, and the operation of the transmission system, in particular congestion management, to promote fair competition and economic efﬁciency. The second directive represents a much more prescriptive and coherent attempt to stimulate progress toward a single electricity market through a top-down approach. However, it quickly became evident to the commission that the problems inherent in moving from theory to practice were no fewer with the second directive than the ﬁrst, and that important gaps remained.

EBL

3.3.2. Regional harmonization A 2004 report31 from the commission assessed and commented on progress in a number of aspects of harmonization, and produced a less than healthy prognosis of progress with implementation. This commentary raised issues concerning the appropriate way of judging progress and indeed how success or failure is to be assessed. The 2005 benchmarking report, the most recent, also hinted that a different approach was needed: The most important persisting shortcoming is the lack of integration between national markets. Key indicators in this respect are the absence of price convergence across the EU and the low level of cross-border trade. This is generally due to the existence of barriers to entry, inadequate use of existing infrastructure and – in the case of electricity – insufﬁcient interconnection between many member states, leading to 29

Regulation 1228/2003, dated 15 July 2003. http://eur-lex.europa.eu/smartapi/cgi/sga_doc?smartapi! elexdoc!prod!CELEXnumdoc&lg=EN&numdoc=32003R1228&model=guicheti 30 That is, the commission may only amend the guidelines following approval by a committee of member state representatives. 31 2004 Report of the Commission on Implementation of the Gas and Electricity Internal Market, at http://ec.europa.eu/energy/electricity/publications/doc/2004_07_09_memo_en.pdf

Achieving Electricity Market Integration in Europe

115

congestion. Moreover, many national markets display a high degree of concentration of the industry, impeding the development of effective competition.32 Furthermore, the report continues, “it must be recalled that the objective of opening the market is to create a single electricity and gas market, not a juxtaposition of 25 national markets. This is a challenging task and integration of all national markets will not happen overnight. At the moment the degree of market integration remains insufﬁcient.” To address problems of this nature, the commission set out a statement of the vision and process for the creation of a single electricity market in its March 2004 strategy paper Medium-term Vision for the Internal Electricity Market.33 This important statement anticipates for the ﬁrst time the integrated single market being reached via the interim step of the establishment and development of a number of regional markets. In turn the commission’s more limited focus at least over the near term was that more compatible arrangements even if applied locally could lead to the application of similar rules for market players regionally and hence promote efﬁcient trade. In other words there was an acceptance that the objective of a single harmonized energy market cannot be achieved even if (big if) all member states were to implement the relevant legislation. The overall goal of the EU vision remains the same – for the market to function as a single market – but the means to this end is somewhat different. Eventually, all TSOs should use to the greatest extent possible the same mechanisms and approaches to manage their networks, and network users should face a single, consistent interface. Thus, over the longer term, the commission still believes a pan-European tariﬁcation mechanism may contribute to the further integration of markets. For the foreseeable future, however, the “vision” paper outlined an approach whereby tariffs for cross-border trade are a combination of different national tariff schemes and where TSOs are compensated for transit and/or other costs and inducing ﬂows is the “most sensible” way forward. For congestion management and system operation more generally, it considers that methods for allocating capacity should be market-based and designed to give correct locational signals to producers and consumers. This approach should help regulators and participants to identify appropriate interconnection projects, depending on the volatility of the signals. Congestion management methods should also be non-discriminatory so that all participants should have an equal chance of obtaining capacity for both long-term and short-term transactions. There should also be an automatically functioning use-it-or-lose-it rule, with mechanisms for reallocation in the event of unused access rights. These objectives imply coordination of the congestion management process with that of short-term energy trading, including the intraday and balancing markets, as well as ancillary services. Finally, such harmonization efforts imply a review of network security rules, grid codes, and access and tariff methodologies, such that trade within a region is no more difﬁcult than trade within a TSO control area. The commission goes on to say it believes that difference of treatment between internal and cross-border congestions was an essential ﬂaw of the Californian market model and that this characteristic is potentially present in some regional European markets. It believes this issue can ultimately be solved by adoption

EBL

32

http://ec.europa.eu/energy/electricity/report_2005/doc/2005_report_en.pdf. The commission issued benchmarking reports in 2001, 2003, 2004, and 2005. A 2006 report is listed on the commission’s website but this is in fact the energy sector package issued in January 2007, which is covered in more detail below. 33 http://ec.europa.eu/energy/electricity/ﬂorence/doc/ﬂorence_10/strategy_paper/strategy_paper_ march_2004.pdf.

Competitive Electricity Markets

116

of a nodal pricing model coupled with ﬁnancial transmission rights, as proposed by the Federal Energy Regulatory Commission’s (FERC) Standard Market Design or competing models. According to the 2004 vision paper, the following speciﬁc objectives are to be pursued. In the medium term: • • • •

• •

ITC should be allowed for suitable compensation between member states for, as a minimum, transit ﬂows and other cross-border ﬂows in some cases; transmission charges on generators should be harmonized within a fairly narrow range with, if appropriate, some locational signals introduced at the EU level; interconnection capacity should be allocated by non-discriminatory, market-based mechanisms consisting of either; within regional markets, there should be a single common coordinated market-based mechanism that allows for both “market coupling” encompassing existing dayahead and possibly intraday spot markets via the adoption of a common timetable, as well as long-term ﬁnancial hedging; between regional markets, there should be speciﬁc market-based mechanisms that, as far as possible, allow for coupling of wholesale markets; and there should be a high degree of transparency to network users, including the publication of necessary data relating to transport capabilities of interconnector lines.

In the longer term: •

there should be both tariffs and ITC based on a single common model of the European network with, ultimately, zone-based transmission charges at the EU level covering, as a minimum, losses, and also potentially ﬁxed investment costs; and • there should be regional market areas served by a single wholesale market (allowing both day-ahead and within-day nomination), which would contain different price areas in the case of persistent congestion.

EBL

In addition to this shopping list the commission called for the following: •

• • •

•

•

transparency, with network users having access to harmonized information on both the transmission networks and the behavior of producers and consumers in the electricity market; mechanisms to allocate capacity designed to avoid manipulation by, and/or collusion of, those generators with a dominant position in certain regions; effective monitoring of market behavior; where congestion management is based on short-term wholesale markets, participants should also have the scope to make longer-term arrangements to obtain ﬁnancial certainty for longer-term contracts through the use of hedging instruments; system operation between national markets should be fully coordinated, and crossborder capacity should be increased through re-dispatching within the national network and through the separation of national markets into different prices areas; and use should be made of revenues from congestion or any other constraint resolution scheme, and these should be clearly regulated and audited, and preferably used for the removal of congestion.

To date a number of key milestones have been achieved or are proposed consistent with this blueprint outlined in Box 3.3.

Achieving Electricity Market Integration in Europe

Box 3.3 Commission medium-term vision – Key milestones Achieved milestones have been: 2004 •

entry into force of the cross-border trade regulation, adoption of ﬁrst set of guidelines, for entry into force on 1 January 2005 including; • a pilot project on co-ordinated congestion management in the “West European market region” was undertaken. 2005 • • •

agreed methodology for inter-TSO compensation came into force; agreed rules for initial harmonization of transmission tariffs were adopted; there was introduction of congestion management based on non-discriminatory and coordinated market-based mechanisms for all congested interconnectors; • UCTE handbook providing core common parameters for system operations became legally binding; • TSOs were required to provide an audited report on amount and use of congestion proceeds and reconciliation; • TSO unbundled accounts implemented.

EBL

2006 •

market participants in all member states gained access to a relevant functioning power exchange by this date; • congestion management methods coordinated between member states for market coupling by the power exchanges concerned within regional market areas enabled; • feasibility study on integration of balancing mechanisms initiated. Looking forward, further, key targets are: 2008 • • •

review of inter-TSO compensation mechanism; possible introduction of locational signals for generation at European level; introduction of regionalized wholesale markets.

2008–10 •

integration of intra-day and balancing markets.

2010 onwards •

progressive integration of all regional groupings.

117

118

Competitive Electricity Markets

3.3.2.1. Ergeg In the absence of sufﬁcient progress toward a single electricity market, the commission and the national regulators are looking to consolidate best practices and pursue a more devolved path to an integrated market. In deﬁning this process it has set out a much more involved role for the regulators. Consequently, an intermediate solution – and an essential transitional step – is to regionally organize physical balancing and market rules across control zones by 2010. Thereafter the aim will be to integrate regional groupings beyond that date. To this end a series of regional initiatives are being supported by the commission. To bring better engagement in the process from national regulators, the commission has explicitly tied them into the process. The European Regulators’ Group for Electricity and Gas (Ergeg), which the commission set up on 11 November 2003 by Decision 2003/796/EC, is an advisory group of independent national regulatory authorities established to assist it in consolidating the internal market for electricity and gas. Its members are the heads of the national energy regulators in the member states. There is considerable overlap with the work and membership of CEER, but Ergeg has a speciﬁc mandate to advise the commission and is formally constituted under community law.34 Ergeg’s work is led by three focus groups, dealing separately with electricity, gas, and transparency. The Electricity Focus Group is developing views and policies on priority topics. The focus group has taken as its broad agenda the provisions of regulation 1228/2003. The commission in 2004 asked Ergeg to consider these topics and advise it concerning texts for the three guidelines envisaged by the regulation. To do this, it set up three task forces under the auspices of the Electricity Focus Group. They are: •

EBL

System Operation Task Force – this deals with security and reliability issues, guidelines on good practice for electricity balancing markets integration, the congestion management guidelines, ﬁrm transmission access rights, and coordination between TSOs; • Cross-border Trade Task Force – this leads work associated with transmission tariff harmonization and guidelines for a longer-term ITC mechanism; and • Regional Electricity Markets Task Force – Ergeg’s launch of the Electricity Regional Initiative (ERI) in spring 2006 and its monitoring and reporting is a key part of the Ergeg’s work program, as we explain in Section 3.4. This task force manages this workstream. Key issues within its locus are reviewing how national legal and commercial conditions for cross-border electricity trade can create barriers to crossborder trade and the framework for cross-border transmission investment. Much of the work of the ﬁrst two groups has developed out of pre-existing initiatives led by ETSO and the commission, though the regional initiative is wholly new (Box 3.3). 3.3.2.2. Electricity regional initiatives Following the 2005 public consultation The Creation of Regional Electricity Markets, which reiterated support for a regional approach to market integration35 and made a concerted 34 Ergeg and CEER have no formal powers, and the commission has been trying to establish a single national regulator for wholesale electricity trades between states, closer to the US model of FERC. To date, member states have been largely unenthusiastic about this direction. 35 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_DOCS/ERGEG_DOCUMENTS_ NEW/ELECTRICITY_FOCUS_GROUP/ERGEG_CREATION_OF_REM_%20DISCUSSIONPAPER_ PUBLICCONSULT.PDF

Achieving Electricity Market Integration in Europe

119

Fig. 3.5. Regional electricity markets. Source: Ergeg 2006.

call at the 12th Florence Forum in September 2005, Ergeg announced a major EU-wide initiative of electricity Regional Energy Markets (REMs) in Spring 2006. The workstream was launched with the publication of its conclusions paper, which conﬁrmed action areas and set out a proposal for the ERI.36 It created seven electricity regions as an interim step (the ﬁrst of many) toward a fully functioning single electricity market, as illustrated in Fig. 3.5. Each region is tasked with identifying and seeking to remove the barriers to competition in its region and report, via Ergeg, to the commission and the Florence Forum. Each regional initiative is intended to bring together regulators, companies, governments, the commission, and other interested parties. The seven regions, some of which overlap, will have similar aims “but reﬂect the very different concerns and level of progress so far achieved in each region.” The process was initiated through a series of mini-fora in spring 2006, which brought together local players under the aegis of the participating national regulators and TSOs. The common task set for the seven groups include:

EBL

• • •

establishing of functioning regional markets; identifying and publishing of priority areas; establishing and publishing an action timetable; identifying who should take action; and • monitoring and fostering progress. In establishing priority topics a number of common themes have emerged from the regional groupings, including: • •

36

congestion management; interconnections;

www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_DOCS/ERGEG_DOCUMENTS_ NEW/ELECTRICITY_FOCUS_GROUP/ERGEG_REMCREATION-CONCLUSIONS_2006-02-08.PDF. There is a parallel workstream for regional gas markets, which was launched in April 2006, and is based round three regional groupings.

Competitive Electricity Markets

120 • •

balancing; and regulatory interfaces.

The initial focus of the project is to map out action plans and timetables, and effort is then to be directed at tackling the ﬁrst priority topics. The next step will be for the regional workstreams to identify a second raft of priority topics. An overall monitoring process overseen by Ergeg is intended to ensure that progress toward a single energy market is not hampered by the regional initiatives, and that each regional initiative is producing proposals and solutions that are compatible with each other. Given the overall goal of a single market for electricity, Ergeg will therefore retain a strong role in overseeing and comparing progress in each of the regional initiatives. Ergeg will for example ensure the cooperation, coherence and compatibility of relevant developments among the different initiatives, relying on Ergeg regular work and procedures as well as on any necessary dedicated tasks.37 There is also a Regional Co-ordination Committee, together with the Implementation and Stakeholder Groups. The TSOs are integral to this initiative and its prospects for success. A lead regulator has been appointed for each REM to chair and coordinate work within the region. Ergeg will report at meetings of the Florence Forum on progress. Each REM must also provide quarterly reports “about measurable results and progress,” 38 primarily as an internal monitoring tool, “to illustrate that the ERI is producing concrete results and delivering on the expectations raised.” The reports are to be accompanied by quarterly

EBL

FLORENCE FORUM EU Commission

Overall forum for discussing progress in progressing regional initiatives and panEuropean consistency

ERGEG

MEMBER STATES

established Regional Co-ordination Committees (RCCs). Each RCC: – comprises regulators from the region; – has autonomy, & adopts its own decision making process; & chairs minifora

IMPLEMENTATION GROUP Includes TSOs, Market Operators (where appropriate) This is the ‘do-er’ for each Regional Initiative Chaired by RCC

Input into identifying and solving market integration issues – e.g. legislative proposals, Member State to Member State interface

STAKEHOLDER GROUP (Based on mini-fora) Wide participation including market players Each mini-fora chaired by RCC and used for consultation

Fig. 3.6. Regional initiatives project structure. Source: Ergeg 2007 work plan.

37

Link as above, p. 39. www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_DOCS/ERGEG_WORK_ PROGRAMME/C06-WPDC-06-04_CEER-ERGEG_WP2007_Public.pdf 38

Achieving Electricity Market Integration in Europe

121

workshops. These reports will be used in turn to produce an annual report, submitted which was39 to the 14th Florence Forum scheduled for September 2007. They will also be used to produce a “convergence and coherence” review report, an initial assessment of the seven REMs, also for discussion at Florence. Based on these outputs the Electricity Task Force will then begin work on a further strategic vision paper, outlining its view on the features that a single market will need to have, plus a road map on how to proceed from REMs to “a true single market.” Ergeg hopes to ﬁnalize this paper during 2008. The structure for rollout of the regional initiatives is shown in Fig. 3.6. Section 3.4 will review the main features of this diagram in more detail. 3.4. Removing Barriers to Market Harmonization This section considers various European-level initiatives that have occurred to stimulate cross-border trade and to help move toward the single electricity market. In several instances the work programs were initiated stimulated by the TSOs acting jointly through ETSO, but over recent years the commission has taken over leadership. The reporting mechanism through the Florence Forum has remained a key means of achieving sign-off of policy development and its outcomes. More recently the national regulators have begun to take ownership of some workstreams and develop their own complementary streams. Eight speciﬁc areas of work are described in this section: • • • • • • • •

managing system-to-system costs; transmission constraints; transmission tariﬁcation; cross-border investment framework; balancing management; information transparency and management; regulatory independence and powers; and functional unbundling.

EBL

While these are the main workstreams they are not the only ones.40 Further, each area is tending to converge and integrate with the regional initiatives. The section also brieﬂy reviews the main points from the energy sector review initiated in 2005, which reported in January 2007, in so far as it addresses impediments to market harmonization. 3.4.1. Managing system-to-system costs The tariﬁcation of cross-border network access between member states has been on the priority list for consideration since liberalization commenced at the EU level in 1996 and has been an active area of focus since 2001, when ETSO published proposals for a temporary CBT. A workstream under the auspices of the commission evolved from this, and a task force was organized and assigned with identifying an ITC mechanism. 39

http://ec.europa.eu/energy/electricity/ﬂorence/doc/ﬂorence_14/ergeg_eri_r.pdf Under the aegis of Ergeg further streams have recently been initiated for electricity on technical issues such as voltage quality and retail competition issues on best practice propositions dealing with customer protection, the supplier switching process, and transparency of prices. 40

122

Competitive Electricity Markets

This is a key mechanism to enable elimination to barriers to trading as the intention is that exchanges across-borders should be treated as far as possible free of barriers when compared to trade within borders, but recognizing that cross-border ﬂows do affect neighboring networks and create costs. “The ITC mechanism is designed to deal with the costs associated with this effect and compensate network owners. By identifying the cost TSOs incur due to cross-border PX, and by compensating them, the ITC mechanism allows trade to function as if there were no such cost.”41 Since 2002 ETSO members have used a temporary system to address the costs of hosting cross-border ﬂows. Initially the focus was on designing an ITC mechanism, and how the mechanism implemented in 2002 should work. ETSO has developed a model to deﬁne the horizontal network of each country. The costs associated with transits (including network losses) on the identiﬁed assets are determined, and the level of ITC is then calculated with appropriate funding set aside into a reserve fund.42 In 2006 the fund was E395 mn. Settlement payments are initially calculated using prior year data but then adjusted for actual power ﬂows. More recently the development has concentrated on enlarging the area of countries among which ITC payments are made43 under the joint mechanism and on the structure of payments. To date the methodology used to determine the amount of payments has changed little even though the mechanism has been regarded as a preliminary solution.44 This system has now been in use for 4 years. The Ergeg Regional Electricity Markets Task Force has been working with the commission and ETSO and aimed to have a longerterm ITC mechanism in place for 2007. However agreement was not forthcoming, and a further temporary scheme was agreed and implemented in June 2007, and this will run to the end of the year. ETSO announced in December 2006 that no voluntary agreement could be reached so far for a new ITC mechanism. A rollover of the prevailing mechanism to end-January 2007 was instead agreed, with the possibility to extend it beyond that if agreement remained elusive.

EBL

3.4.2. Managing transmission constraints The Florence Forum agreed at its ﬁfth meeting in March 2000 that it was important that congestion management within the union be based on market solutions, which give “proper and justiﬁed incentives to both market parties and TSOs to act in a rational and economic way.” Further work led to an initial set of guidelines for congestion management being set out in the conclusions of the sixth meeting in November 2000. These guidelines emphasized the need for TSOs to set out in a transparent manner the congestion management methods in use, and for TSOs to maximize the available interconnector capacity. They were not legally binding. In 2002 reports on congestion management were produced by, among others, the CEER and the commission, which noted progress but concluded that further work was required. ETSO’s work program on congestion management effectively started in 2001. A position paper was put to the Florence Forum meeting in that year45 , and the organization also 41

http://ec.europa.eu/energy/electricity/publications/doc/2006_03_tso_compensation_mechanism. pdf 42 www.etso-net.org/upload/documents/ETSO%20porposal%20for%202006%20CBT.pdf 43 Britain and the two Irish jurisdictions still do not participate. 44 The only substantive change made to date has been the abandonment of an export fee in 2005. 45 www.etso-net.org/upload/documents/Position_Paper_on_Congestion_Management.pdf

Achieving Electricity Market Integration in Europe

123

considered how it should evaluate congestion management methods46 recognizing in its conclusions that a mix of methods may be appropriate across the union. The commission then published a communication, which called upon ETSO (with others) to develop the guidelines further and to propose common standards on “soft measures” to increase cross-border transmission capacity.47 This request led to a vision paper circulated in April 200248 and development of an outline of a coordinated congestion management scheme in September that year.49 A methodology for deﬁning and quantifying transfer capacity deﬁnition was subsequently agreed in June 2004.50 ETSO also developed its own position based on ﬂow-based market coupling51 , and it began to address the development of transmission risk-hedging products52 . The Directive 1228/2003 speciﬁcally addressed the need to stimulate convergence among member states in the area of congestion management. General principles are set out in Article 6. Congestion problems must, for example, be solved using market-based solutions that give efﬁcient economic signals. The supporting regulation allows the commission working within a comitology53 process to set binding guidelines on congestion management. The adopted version of the regulation already contained draft guidelines that cover some general aspects of congestion management and allocation of available transfer capacity of interconnections between national systems. It was intended that these be reﬁned by regulators and the commission following consultation. Subsequently Ergeg has taken up the development process, and it issued reworked draft congestion management guidelines in September 2004. and a revised set in 200554 . These entered the comitology process later that year. The cross-border electricity committee gave a positive opinion on the guidelines in June 2006, and the European Parliament then had the right of scrutiny until end-September. Since 1 January 2007 the guidelines55 require use of a common ﬂow-based network model, application of a market-based allocation system for securing rights on transmission capacity, and the use of common products (including conditions, timeframes). The commission in parallel is also scrutinizing some existing long-term priority reservations. A number have already been cancelled, and others are likely to be subject to infringement proceedings.

EBL

46

www.etso-net.org/upload/documents/Evaluation_of_Congestion_Management_Methods_for_ Cross-Border_Transmission_(Florence).pdf 47 European Energy Infrastructure, Communication from the Commission to the Council and the European Parliament, January 2001. 48 www.etso-net.org/upload/documents/ETSO%20Vision%20on%20CCM.pdf 49 www.etso-net.org/upload/documents/Outline%20Proposals%20for%20CCM5.pdf 50 www.etso-net.org/upload/documents/Transfer%20Capacity%20Deﬁnitions%20-%20Final.pdf 51 www.etso-net.org/upload/documents/ETSO-EuroPEX_Interimreport_Sept-2004-.pdf 52 www.etso-net.org/upload/documents/Short%20ETSO%20Risk%20hedging%20in%20CM_ﬁnal%20 PUBLIC.pdf 53 That is, the commission may only amend the guidelines following approval by a committee of member state representatives. [Moved] 54 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_PC/ARCHIVE1/CM_GL/ERGEG_ AMENDMENTS_GUIDELINES_CM_PUBLIC_CONSULTATION_20050.PDF 55 http://ec.europa.eu/energy/electricity/ﬂorence/doc/ﬂorence_13/cm_guidelines.pdf

124

Competitive Electricity Markets

3.4.3. Transmission tariff harmonization Progress in this area has also been intermittent. At the ﬁfth meeting of the Florence Forum in March 2000 parties stressed the importance of making progress in harmonizing the split of transmission charges between generation and load across member states. ETSO initially brought some focus in 2002 when it formed a task force to consider transmission tariﬁcation methodologies and charges across member states, and it issued a benchmarking study based on 2002 tariffs of 12 member states56 and again in 2005 based on 2004 tariffs of 19 member states.57 A legal basis for dealing with the issue came with the adoption in June 2003 of Regulation 1228/2003. In particular: •

Article 8(3) provided for guidelines to be adopted to “determine appropriate rules leading to a progressive harmonisation of the underlying principles for the setting of charges applied to producers and consumers (load) under national tariff systems, including the reﬂection of the ITC mechanism in national network charges and the provision of appropriate and efﬁcient locational signals, in accordance with the principles set out in Article 4”; • Article 4 discusses the requirements relating to transmission tariffs. In particular, Article 4(2) states that “where appropriate, the level of the tariffs applied to producers and/or consumers shall provide locational signals at European level, and take into account the amount of network losses and congestion caused, and investment costs for infrastructure”; and • Article 4(4) requires that providing that appropriate and efﬁcient locational signals are in place, charges for access to networks applied to producers and consumers shall be applied regardless of the countries of destination and origin, respectively, of the electricity, as speciﬁed in the underlying commercial arrangement.

EBL

The Regulation allows the commission, again working within a comitology process, to set binding guidelines. In June 2003, the commission produced a discussion document on harmonization of network access charges.58 This consultation provided the ﬁrst basis for discussion of form and content of guidelines. The subject was discussed further during the 10th Florence Forum of July 2003, and the commission issued draft guidelines in March 2004. Subsequently Ergeg and the commission have worked together to produce a reﬁned text, which was presented to the 11th meeting of the Florence Forum in September 2004. Ergeg further worked up the text, and it formally consulted on this in May 2005.59 It then gave its advice to the commission on tariﬁcation guidelines in July 2005. The objective of the proposed guidelines was the ﬁrst step to harmonize charges paid by generation. The commission formally put a proposal on tariff guidelines to the comitology procedure in July 2006. Ergeg approved this in December 2006, and also stated that it considered that there was a need for widening the tariff harmonization to lower voltage levels. It also approved the details for calculating and reporting of transmission tariff levels under the 56

www.etso-net.org/upload/documents/Benchmarking.pdf www.etso-net.org/upload/documents/08-04-05%20Synthesis%202004%20FINAL%20%20.pdf 58 www.europa.eu.int/comm/energy/electricity/ﬂorence/doc/ﬂorence_10/g_and_l/ec_g_ harmonization.pdf 59 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_PC/ARCHIVE1/TT_GL/ INTRODUCTORY_NOTE_ERGEG_AMENDMENTS_GUIDELINES_TARIFICATI.PDF 57

Achieving Electricity Market Integration in Europe

125

proposed tariﬁcation guidelines. The intention was that the guideline will come into force at the same point as the ITC guidelines. Ergeg was preparing the ﬁrst report on transmission charging structure and generator values during the ﬁrst half of 2007. Furthermore, it continues to address the issue of tariff harmonization in its work program for 2007, as part of which more studies on the transmission tariff structures and further harmonization are planned. 3.4.4. Cross-border interconnector framework Cross-border investment has also been part of a wider series of initiatives by the commission. They were emphasized in the Green Paper, a European Strategy for Sustainable, Competitive and Secure Energy,60 in which the commission noted there that “there is an urgent need for investment” in electricity sector infrastructure, especially interconnection between states. It also stated that private and public investments in infrastructure needed to be stimulated, and called for authorization procedures to be accelerated. Consequently it committed to identify by end-2006 individual measures that it considered important at the member state level. Further, in order to promote EU single internal markets and to foster economic and social cohesion, the treaty establishing the union also provides a legal basis for the further development of Trans-European Networks (TENs), including those for electricity. This development includes the interconnection and interoperability of national transmission networks, as well as access to such networks.61 In January 2006 Directive 89/200562 concerning measures to safeguard security of electricity supply and infrastructure investment was adopted. Among other things, the directive establishes measures aimed at ensuring an appropriate level of interconnection between member states for the development of the internal market. Measures also include provisions related to ensuring operational network security and appropriate network investment, taking into account relevant market actors and TSOs. In addition, member states are also permitted to require TSOs to provide information on investments related to the building of internal lines that materially affect the provision of cross- border interconnection. In support of these initiatives and also of the implementation of the 2003 regulation, Ergeg’s June 2005 public consultation paper The Creation of Regional Electricity Markets re-emphasized cross-border investment as a priority issue for development, and it included the matter in its work program for 2006. The consultation was also partly informed by two papers from ETSO on cross-border investment.63 Ergeg subsequently launched a public consultation on the cross-border framework for transmission investment in October 2006.64 This paper set out views of an Ergeg expert group on the issues, in particular the process

EBL

60

http://europa.eu/scadplus/leg/en/lvb/l27062.htm Accordingly, the European Union provides some ﬁnance under the remit of TENs for some electricity and gas transmission infrastructure projects of European interest. A yearly budget of about E25 mn is spent mainly for supporting feasibility studies. Guidelines on TENs specify which projects are eligible for funding. Financial rules specify the ﬁnancial procedures involved. 62 http://eur-lex.europa.eu/LexUriServ/site/en/oj/2006/l_033/l_03320060204en00220027.pdf. The directive has to be implemented by 24 February 2008. 63 Roles and Responsibilities of TSOs and other Actors in Cross-border Network Investment and Overview of Procedures for Building 110 kV to 400 kV lines, which both called for the expedition of planning processes. 64 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_PC/Cross%20Border%20 Framework/ERGEG_CrossBorderFramework_PC_2006-10-04.doc 61

126

Competitive Electricity Markets

for obtaining building and construction authorizations and permissions for transmission network infrastructure, which “needs to be recognized and tackled.” At present provision of transmission network infrastructure is largely driven by national law and requirements and “the obligations on authorities or TSOs seldom or insufﬁciently extend to cross-border infrastructure or the need to integrate markets.” The paper makes recommendations regarding an appropriate framework for the provision of cross-border transmission infrastructure. In particular Ergeg called for: •

extensions to planning and operation standards on TSOs to include cross-border obligations; • adjustment of regulators’ duties and competences to include some cross-border or regional elements; • reorientation of the role of TSOs in order that they act collectively, subject to regulators; and • views on the feasibility of a merchant model for development of new interconnectors and its regulatory treatment. In support of this effort ETSO has established a public database on interconnections between participating countries. In addition to physical details and scheduled outage data, there is also data on allocation mechanisms and results of their application. It is also being expanded to include common data on grid availability as well as national load and generation data. As ETSO makes clear, “The responsibility for data edition, checking and publication is clearly on individual TSOs. For many legal and organisational reasons, ETSO cannot act in their name. But ETSO may make the information access easier. It can also allow a quick overview of what is available, and of the reasons why a piece of information may not be available in a given area.”65 In support of its work in this area and to increase market transparency ETSO has also recently launched its “ETSO-vista platform,” which sets out hourly power ﬂows between national markets.

EBL

3.4.5. Balancing services ETSO has initiated a process that is leading to scoping the issues associated with balance management and enabling cooperation between national operators and markets. As a ﬁrst step in December 2003 it formed a task force that produced a report intended to increase understanding of prevailing current approaches to balance management, which it described as “the management processes and services associated with power system operation, which ensure quality and short-term security of supply (but in particular focusing on active power (MW) balancing and frequency control.”66 The report observed that there are a number of fundamental physical differences between transmission systems across Europe, and these affect the way in which balance management is approached currently (and could be approached in the future). These differences include the size and inertia of the system, AC or DC interconnection, availability of secondary control, and type of generating plant connected. A key conclusion of the task force was that different national labels concealed the procurement and application of essentially the same services and products. The survey 65 66

www.etso-net.org/MarketInfo/marketdata/e_default.asp www.etso-net.org/upload/documents/BalanceManagemeninEurope.pdf

Achieving Electricity Market Integration in Europe

127

identiﬁed a number of common features, but there were a variety of ways of paying for services. The scope of reserves available to TSOs is varied and wide. The limits imposed on the TSO in terms of dispatching reserve in a balancing mechanism also vary. All reserves and energy balancing services are procured on a commercial basis. The payment mechanism for such services tended to be based upon a utilization payment for energy and an availability payment based on capacity, of which there are different types depending on contractual arrangements and the type of service. Primary control reserves for the individual groups UCTE, Nordel, and England and Wales and Scotland are generally shared between TSOs based upon relative annual consumption or generation. Secondary control only applies to countries in UCTE, and the TSO responsible for managing a control block coordinates the control area TSOs in that block. A number of different arrangements were described for gate closure (when expected physical notiﬁcations become ﬁrm). In some countries this is day-ahead, some are ﬁxed windows in the operational day and some are based upon a rolling settlement period basis. There are a range of times between gate closure and real time in use, including 1 minute, 1 hour, 3 hours, and day-ahead. Calling services via a balancing mechanism gives rise to two common types of pricing. These are either “pay as bid” (most countries) or marginal pricing (the Netherlands, Sweden, Norway, Finland, Spain, and Greece). In November 2005 ETSO produced a further report, which explored the prospects for establishing or increasing cross-border trading of balancing services, and which examined the commercial market design and relevant technical issues associated with facilitating an increase in such activity. “The potential beneﬁts of increasing opportunities for TSO balancing activities across-borders relate primarily to the procurement of balancing services in an efﬁcient manner by the possibility of using internal and external reserve resources in a competitive environment across a wider area,” the report noted, adding “this is likely to have associated beneﬁts in relation to plant loading efﬁciency (with associated environmental beneﬁt) and a decrease in balancing costs, which are ultimately borne by consumers.”67 The task force went on to develop two generic models that describe how trading of balancing services across borders could be realized, including the principles of the models, covering the roles and responsibilities of each participant. The two conceptual models that have been developed are: (i) reserve provider to TSO trading and (ii) TSO to TSO trading. It also addressed how the models function in practice based on case studies and documented the issues associated with each of them. However, the task force emphasized that it is not intending to restrict possible solutions to these two models. A third report was issued in May 2006.68 It examined the commercial market design and relevant technical issues, including security of supply, associated with facilitating an increase in cross-border trading of tertiary (i.e., manually instructed) reserves for system balancing purposes. More speciﬁcally, ETSO’s objective in this work stream was “to analyse the consequences of steps in the integration of these markets if the differences (or incompatibilities) in balancing mechanisms are kept, in order to identify measures needed to allow these integration steps.” It examined the potential beneﬁts of increasing opportunities for TSO balancing activities across borders “relating especially to the procurement

EBL

67 www.etso-net.org/upload/documents/Report%20BMnov2005.pdf. The report went on to note: “The issues and potential beneﬁts are similar to those that could result from increased facilitation of energy trading between markets, particularly in respect of intra-day energy trading. Many of the issues associated with the facilitation of energy trading relate to the harmonisation of market and trading rules across Europe, which will be addressed in later work.” 68 www.etso-net.org/upload/documents/Report%20BM%2018-05-06.pdf

128

Competitive Electricity Markets

of tertiary reserves through use of both internal and external reserve resources in a competitive environment across a wider area, to achieve better plant loading efﬁciency and a decrease in balancing costs.” The trading of automatically delivered frequency control services (primary and secondary control) was intentionally not considered due to the much more complex nature of these services. Additionally, “it is believed that there is potentially more economic value in developing arrangements related to cross-border trade of tertiary reserves and balancing energy for timeframes between day ahead and some minutes before real time (that is before the action of automatic devices),” the report notes. The analysis examined four models (and several variants) for the procurement and trade of tertiary reserves, the impact of ﬁnite cross-border transmission capacity, and the possible need for possible priority rules for these services. It observed that different TSO roles lead to different balancing mechanisms, which use incompatible products, but that differences in gate closure times across balancing markets in Europe do not have a strong impact on the possibility of cross-border balancing services trading, but that differences in bidding rules do. It advocated that differences in regulatory regimes regarding reserve markets should be leveled out to prevent buying out of reserves from areas with a low reserve price. The next assignment for this task force is to develop a balance management reference model that can be a “guideline in the development process towards facilitation of crossborder tertiary reserves trade and regionally integrated balancing mechanisms.” Ergeg has also been developing views on balancing convergence and is developing guidelines on good practice in this area following the presentation to the 12th Florence Forum in September 2005 of its position on balancing mechanisms compatibility. Separately the commission appointed consultants Frontier Economics and Consentec to look at ways in which trade close to real time could be facilitated and to recommend measures that would help to improve the efﬁciency of cross-border trade “without imposing a disproportionate cost on the systems in question,” which has informed the Ergeg process.69 Following consultation, Ergeg adopted initial advice to the commission on the aspects of electricity balancing markets integration in line with the congestion management guidelines in December 2006,70 and which constitutes its initial advice to the commission on electricity balancing markets integration as required by the 2003 regulation. Final advice is to be provided to the commission after the development and consideration of the aspects on intraday markets and automatically activated reserves as part of the 2007 work program.

EBL

3.4.6. Market transparency In March 2006 Ergeg launched a public consultation on guidelines for good practice on information management and transparency in electricity markets. These guidelines 69

http://ec.europa.eu/energy/electricity/publications/doc/frontier_consentec_balancing_dec_ 2005.pdf 70 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_DOCS/ERGEG_DOCUMENTS_ NEW/ELECTRICITY_FOCUS_GROUP/E05-ESO-06-08_GGP-EBMI_2006-12-06.pdf Beneﬁts and Practical Developments of Intraday Electricity Markets and Balancing Mechanisms, December 2005. The report looked at three basic types of integration: linkages of intraday markets (i.e., before either system gate closure); linkages of balancing arrangements after one gate closure; and linkages of balancing arrangements after both gate closures. It concluded that, given that in many cases model 1 may be capable of implementation relatively quickly, a phased approach to the development of ﬁnal arrangements using model 1 as a transitional step within regional blocks may be worthy of consideration. Such a phased approach may have the advantage that, under model 1, harmonization between balancing regimes is less important – a number of differences can be internalized by the TSOs.

Achieving Electricity Market Integration in Europe

129

seek to establish a consistent approach to the provision of market-related information by wholesale market participants across member states.71 The guidelines set out Ergeg’s views on the minimum required level of transparency across the European market. Moreover, they are intended to give a set of rules required for the organization of information and its dissemination across the markets of common member states. Finally, they deﬁne general principles governing information release, either through publication or through information released to market participants on request. Unlike other areas where guidelines are anticipated by the 2003 regulation, this initiative has developed from further thinking by the national regulators themselves. The issue of transparency has been one of increasing concern at Florence Forum meetings, and at the 12th forum a number of stakeholders (including Eurelectric,72 EFET,73 ETSO74 ) stressed the need for greater market transparency. Ergeg said in its consultation that it would “anticipate that regulators, working together with stakeholders, will put in place and monitor functioning of the guidelines to the extent possible. Such work is likely to involve dialogue and discussion with TSOs and market participants. 2007 will therefore be a year of appraisal.”75 In practice other legislative developments are likely to have a bearing on this period of monitoring and appraisal. The ﬁnal reports were issued in January 2007, and both contained analyses and remedies extending to transparency issues, which are now being assimilated. Market integration through the formation and continuation of the regional initiatives proposed by Ergeg is also likely to raise strongly the issue of transparency, and lead to further development of policy. 3.4.7. Unbundling

EBL

Article 19 of Directive 2003/54/EC set out the legal minimum requirements of accounting unbundling. However, the commission’s 2005 benchmarking report showed that insufﬁcient unbundling was one of the key factors reducing retail competition in many markets. It observed that the development of wholesale markets was impeded by too close a relationship between TSOs and afﬁliated producers or importers. Ergeg launched a public consultation on guidelines for good practice on regulatory accounts unbundling in April 2006. The guidelines which were issued in April 2007, establish basic principles on regulatory account unbundling, reinforcing the requirements of the second directive, under which by 1 July 2007 distribution network operators with more than 100,000 customers were required to be unbundled in legal terms. The guidelines recognize the importance of TSO independence; separation of distribution and supply businesses; and other separation requirements. The guidelines are directed to regulators, but the group hopes they can also serve as a benchmark for analyzing national unbundling of accounts. The presented guidelines are the ﬁrst set of unbundling guidelines. Those on informational and management unbundling will be elaborated after more experience is gathered on the actual implementation of the relevant requirements. 71

Information management and transparency for information to retail customers is the subject of separate work being undertaken by Ergeg with the CEER. 72 http://ec.europa.eu/energy/electricity/ﬂorence/doc/ﬂorence_12/eurelectric_iem.pdf. Eurelectric is a trade association of power utilities. EFET is the European Federation of Energy Traders. 73 http://ec.europa.eu/energy/electricity/ﬂorence/doc/ﬂorence_12/efet_iem_p.pdf 74 http://ec.europa.eu/energy/electricity/ﬂorence/doc/ﬂorence_12/etso_iem.pdf 75 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_PC/ARCHIVE1/GGP_ Transparency/E05-EMK-06-10_GGPa_transparency_info_mgmt_off.pdf

130

Competitive Electricity Markets

The 2005 benchmarking report again highlighted insufﬁcient functional unbundling as “a persistent impediment” to competition in EU electricity and gas markets. Ergeg criticized what it termed the “vague” implementation of the unbundling provisions of existing directives so that when they are transposed into national law they leave a lot of discretion to the integrated companies who “often act without fear of enforcement because of a lack of adequate powers of the regulators.” To address this issue, Ergeg recommended that new legislation is required. Further, the report says regulators need to have the appropriate powers to ensure wholesale and generation market transparency and that information and data held by network operators is either ring-fenced or released to the market in a non-discriminatory manner. It also recommends that regulators monitor and share information with each other on cross-border unbundling arrangements and develop and/or enforce binding guidelines on unbundling where necessary. The sector inquiry has underlined the commission’s concerns in these areas. 3.4.8. Regulatory independence and powers In its 2006 review Ergeg identiﬁes the gaps between what it terms “the reality and the desired set of competencies of regulators,” and makes recommendations on how to close these gaps.76 The report says that while all national regulators are separate from the relevant ministries in charge of electricity, those in Austria, France, Germany, Greece, Italy, and Malta still retain some powers to approve, reject, or amend regulatory decisions. The recommendations were part of a series of reports submitted to the energy commissioner Andries Piebalgs in December, assessing the development of EU energy markets in 2006. Ergeg called for more legislation to provide national regulators with the power to: • •

EBL

facilitate the development of and competition in the internal electricity market; enable them to act jointly to oversee investment in transmission networks to ensure appropriate cross-border transmission capacity; and • enable them to gather and exchange information and pursue monitoring or investigations into activities that occur in one territory that affect markets in another. The assessment also said regulators were concerned of a “real danger” that political declarations of support for regional integration were not being followed through with the necessary action on the ground. “The tendency to put independent regulators under greater political control and protect ‘national champions’ acts against regional integration and the single market objective,” it argued. “Political support to the regional initiatives was also required,” Ergeg added. As far as national markets were concerned, Ergeg said the regulator must be independent of industry and be able to make decisions without government intervention. The regulator must also have a role in advising government on energy markets. In addition, it must have sufﬁcient power to enforce its decisions and apply sanctions and penalties when breaches of market rules occur or there is evidence of discriminatory behavior. The regulator should also be able oversee monopoly network activities. To ensure power markets are “effectively competitive,” the group also said that regulators need to have the power to approve and monitor compliance with market rules for wholesale trading. They should also be able to require market information and data from 76

www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_DOCS/PRESS_RELEASES/PR-0614_ERGEG_AssesmentNatReports_2006-12-08.doc

Achieving Electricity Market Integration in Europe

131

all participants and have the legal right to pass this on to the relevant national and EU competition authorities in the case of an investigation. Further, the regulator should be able to require all participants to publish or make available any information and data in order to improve market transparency and ensure compliance with national and European market rules. Ergeg called for a “ﬂexible legal approach” regarding any new powers for the regulators as a result of its recommendations. These powers, it said, “need to be sufﬁciently ﬂexible in order to allow regulators scope to make timely and efﬁcient adjustments to the regulatory framework that reﬂect current and anticipated market developments.” 3.4.9. Energy sector review Notwithstanding the Ergeg process the commission’s language on the health of the liberalized single market has continued to become more strident. “Meaningful competition does not exist in many member states. Often customers do not have any real possibility of opting for an alternative supplier…. In summary, stakeholders do not yet have a high degree of conﬁdence in the internal market.”77 As a consequence of these shortcomings, the commission initiated in 2005 an inquiry for the gas and electricity sectors under competition law. The sector inquiry carried out in conjunction with country reviews conducted by the commission during 2006 was completed in January 2007. It “unearthed a variety of speciﬁc examples which demonstrate the shortcomings of the existing regulatory framework.” The focus of these was the predominating vertically integrated structure and the informational and strategic advantages enjoyed by dominant companies, including their continuing ability to exert inﬂuence over TSOs where they owned them. More speciﬁcally:

EBL

•

• •

•

•

TSOs, especially when vertically integrated, failed to create conditions conducive to liquid competitive markets – for example by maintaining localized separate balancing zones rather than facilitating the integration at national and cross-border level: TSOs have been slow to act to increase cross-border capacity; there is evidence that both TSOs and regulators tend to be over-oriented to shortterm national concerns rather than be proactive in trying to develop integrated markets; on many issues, certain regulators are constrained in their relations with the industry, lacking the appropriate powers and discretion; some regulators are also subject to continuing strong direction from national governments; and concentrated national markets have tended to encourage regulators to introduce intrusive regulation into wholesale and balancing markets, for example, price caps, which are a strong disincentive to invest.

As a result of these widespread shortcomings, incumbent electricity and gas companies largely maintain their dominant positions on “their” national markets. This situation has led, in the commission’s view, many Member States to retain tight control on the electricity and gas prices charged to end-users. Unfortunately this is often a serious constraint on competition. 77

http://eur-lex.europa.eu/smartapi/cgi/sga_doc?smartapi!celexplus!prod!DocNumber&lg=en& type_doc=COMﬁnal&an_doc=2006&nu_doc=841

132

Competitive Electricity Markets

In response to the inquiry, a number of investigations have been launched against companies in the electricity and gas sectors. The commission has said it also intends to take action to address remaining issues in the following areas: •

Ensuring non-discriminatory access to networks The commission is considering two main avenues for further TSO unbundling measures, with a view to making formal proposals. These are: – fully (ownership) unbundled TSOs: the TSO would both own the transmission assets and operate the network. It would be independently owned, i.e., supply/ generation companies could no longer hold a signiﬁcant stake in the TSOs; and – separate system operators without ownership unbundling: this solution would require separation of system operation from ownership of the assets. The commission notes that the independent system operator (ISO) model would require detailed regulation and permanent regulatory monitoring, implying clearly that it prefers the independent ownership approach. • Improving regulation of network access at national and EU level The commission came to the conclusion that energy regulators need to be strengthened at the national level and have “the required level of discretion to take decisions on all relevant issues.” The commission said regulators needed strong powers over a number of key areas including: (i) all aspects of third-party access to networks, (ii) balancing mechanisms, (iii) market surveillance of, e.g., PXs, (iv) all cross-border issues, (v) consumer protection including any end-user price controls (vi) information gathering, and (vii) sanctions for non-compliance. It proposed a strengthening of the Directives on this basis. •

EBL

Reducing the scope for unfair competition

In addition to contemplating creation of powers to free up sites for new entrants and to force termination of long-term contracts over-hanging the market, the commission intends to introduce binding guidelines for transparency either through new legislation or by modifying the regulation 1228/2003. The advice of Ergeg will be taken in the ﬁrst instance. •

Providing a clear framework for investment by further harmonizing TSO activity

A higher level of technical cooperation between TSOs is proposed, including detailed exchange of information, both in terms of long-term network planning and on a real-time operational basis. The commission suggests that it is doubtful whether this can be achieved in the current framework where both TSOs and regulators are inclined or even obliged to follow a national focus. An enhanced level of TSO coordination would require a new legislative framework. Existing associations of TSOs would be granted an institutional role, with formal obligations and objectives being added to their role, The TSO could be required by the commission or the regulators to report on European grid operation and investment, as well as the development of common technical standards for network security. It could also be responsible for monitoring the developments of networks so as to improve the transmission capacities between member states. Efforts should also be made to have a gradual evolution toward regional system operators, and cross-border system operators could be set up. These would be independently owned and would require additional functional unbundling. 3.5. Establishing the “Western” Regional Market This section looks more closely at how the jurisdictions that comprise the Western market for electricity – Britain, France, and Ireland – are tackling convergence through the ERI.

Achieving Electricity Market Integration in Europe

133

By January 2007 this was the only one of the seven electricity regions to have published substantive consultations. 3.5.1. Project overview The France, Ireland, and UK electricity REM is termed UK and Ireland, which is rather misleading as it includes France. France also participates in three other regional initiatives given its pivotal position within the Western European electricity system. To avoid confusion we have termed this strand of the ERI theWestern regional market. Annual electricity consumption in the France, Ireland, and UK electricity REM is about 780 TWh and the region comprises about 30% of the electricity market within the EU. This regional market includes two of the biggest national economies in the union. The broad task of the regional initiative is to achieve greater integration, which will facilitate regional development and accelerate cooperation and coordination at the European level. Following a Stakeholders’ Group meeting on 14 November 2006, workstreams have been established to take forward priority issues. The four regulators are Ofgem (Great Britain), CRE (France), CER (Republic of Ireland), and Ofreg (Northern Ireland). It is led by Ofgem. 3.5.2. Priority issues The regional regulators have published a series of documents and papers on priority issues, some in collaboration with the four TSOs – EirGrid, Réseau de Transport d’Electricité (RTE), National Grid, and System Operator Northern Ireland (SONI). They consider key issues for addressing to create “a balanced market” across the different national trading regimes, being:

EBL

• • • •

compliance with the congestion management guidelines; coordination of interconnector capacity allocation; enhancing reciprocal access to balancing mechanisms; and wholesale market transparency.

The TSOs have written a paper for each of the four priority issues. A cover note accompanies each document in order where relevant to set out the background to the document and to invite comment. The TSO’s ﬁrst paper, Gap Analysis of Current Compliance with the Congestion Management Guideline,78 concerns an assessment of compliance of present practice on interconnectors with the guideline. The document looks at the impact these guidelines will have on the regional TSOs and the actions they will need to take to fully comply. The TSO’s second paper, Options for Further Development and Coordination of Existing Cross-Border Arrangements,79 builds on the TSO’s ﬁrst paper and puts forward options for dealing with areas where current congestion management practice in France, the United Kingdom, or Ireland does not meet the requirements of the congestion guideline. It seeks views on what aspects of capacity allocation and coordination require most attention. It 78

www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_RI/ERI/UK_Ireland/Papers%20 and%20Written%20Consultation/20061027_Paper%20_IG2-1CMGuidelineGapanalysis_FINAL.doc 79 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_RI/ERI/UK_Ireland/Papers%20 and%20Written%20Consultation/20061031_Paper_IG2-2CrossBorderInitiatives_FINAL.doc

134

Competitive Electricity Markets

also considers whether greater use should be made of “use it or lose it” type provisions in conjunction with current interconnector access rights, and whether such mechanisms would facilitate trade. It also explores how intraday trade between national markets can best be facilitated on interconnectors. The group notes that the liberalization of European electricity markets has increased substantially the need for interconnector capacity. It believes greater interconnection is needed to boost competition entailing more efﬁcient use of the existing three regional links (the North–South interconnector in Ireland, the Northern Ireland–Scotland Moyle link, and the Interconnexion France–Angleterre (IFA)). The regulators want better coordination of capacity allocation, particularly for IFA and Moyle. This change will “encourage trade, reﬂect national market requirements and ensure that market participants can co-ordinate their desire to trade over two or more interconnectors simultaneously.” The TSOs’ third paper, Conceptual Options for achieving Reciprocal Access to Balancing Arrangements across the Region,80 puts forward options for ways in which TSOs and market participants might trade in balancing market services across borders. It looks at current regional TSO mechanisms. Two models are used to show how access may be improved. Both models already exist elsewhere in Europe: the ﬁrst within the German control block and the second within Nordel. One recommendation here is for further development of intraday arrangements on the IFA. More generally, there is a call for improved balancing activities throughout the whole region and between national markets to increase plant loading efﬁciency and to reduce balancing costs. The TSOs’ fourth paper, Situation Report on Wholesale Market Transparency,81 makes an assessment of current levels of wholesale market transparency within the territories. It uses the Ergeg draft Guidelines for Good Practice on Information Management and Transparency as an input for this assessment, and seeks views on what issues raised in the guidelines need to be addressed in the region. Regulators have written further papers concerning imbalance pricing, transmission tariﬁcation, and efﬁciency of interconnectors. The regulators’ ﬁrst paper, Electricity Imbalance Pricing – Comparison of Regimes and Effects across France, GB, Republic of Ireland,82 makes a ﬁrst assessment of the imbalance rules applying in each territory with a view to understanding if signiﬁcant differences exist that might tend to affect trading decisions were full reciprocal access to balancing markets to exist. It complements the TSOs’ third paper. It asks whether the paper addresses the correct issues and whether any existing market distortions are properly described. The issue of imbalance energy pricing is studied by way of a comparison of the different regional regimes. Based on various analyses, this document concludes that trade would be more efﬁcient if imbalance pricing rules in adjacent markets were uniform. This situation arises because any differences between territories might affect the “competitive position of market participants, potentially resulting in inefﬁcient trades.”

EBL

80

www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_RI/ERI/UK_Ireland/Papers%20 and%20Written%20Consultation/20061027_Paper_IG2-3%20Balancing_FINAL.doc 81 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_RI/ERI/UK_Ireland/Papers%20 and%20Written%20Consultation/03112006_Paper%20IG2-4_RTE_Transparency.pdf 82 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_RI/ERI/UK_Ireland/Papers%20 and%20Written%20Consultation/20061101_Paper_IG2-5%20Imbalance%20Pricing_FINAL_.doc

Achieving Electricity Market Integration in Europe

135

The regulators’ second paper, National Transmission Tariﬁcation Systems,83 outlines the current national transmission tariﬁcation systems across the four territories together with a ﬁrst assessment of the extent to which differences in approach might affect cross-border trade. Here, the regulators solicited industry opinion as to whether the systems impact regional trade and, if so, how this can be rectiﬁed, possibly by revising or aligning tariffs in adjacent countries. The regulators’ third paper, Analysis of Interconnector Flows on the English Interconnector Angleterre,84 makes a ﬁrst assessment of the extent to which the electricity interconnector between the markets is used efﬁciently, and puts forward some hypotheses to explain the observed levels of efﬁciency. This paper is intended to complement the ﬁrst and second of the TSOs’ papers. It ﬁnds there is room for greater efﬁciency because there is currently an “irrational” utilization of capacity with some ﬂows occurring in the opposite direction of the price differentials between the two markets. Moreover, the analysis shows that when use of the interconnector has been “rational,” capacity was under-used two-thirds of the time. For this regional market to be successful, the wholesale market will have to be much more transparent in order to facilitate efﬁcient price discovery. Ergeg’s recommendations in this area of information disclosure are based on its draft transparency guidelines. These state that “trade between countries in this region will be best facilitated where the appropriate level of information is made available to market participants.” 3.5.3. Next steps

EBL

These consultations closed in January 2007, and Ergeg and the regional regulators and TSOs are now considering the responses. A detailed project plan was due to be published in March. The focus will be on producing workable proposals for the congestion management, balancing and transparency workstreams, especially development of a model for reciprocal access to balancing markets. The consultations will also inform the further work needed to address the priority issues and to assess what further actions, if any, are needed regarding the three additional topics. 3.6. Conclusions The reform processes to secure electricity market convergence and harmonization in Europe have shown sustained activity over more than a decade. They have in the main been characterized by “more heat than light.” Despite two landmark directives and sustained attempts to impose top-down harmonization, numerous workstreams initially led by technical agencies then by the commission have delivered few tangible or signiﬁcant results, and the commission’s vision of a single internal market remains some years hence. Probably the most notable success has been achieved by the facilitation route, with the promulgation of guidelines under the important 2003 regulation on cross-border electricity exchanges. Four years on this approach is beginning to see real progress in the areas of the 83

www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_RI/ERI/UK_Ireland/Papers%20 and%20Written%20Consultation/20061101_Paper_IG2-6National%20Tarifﬁcation%20Systems_.doc 84 www.ergeg.org/portal/page/portal/ERGEG_HOME/ERGEG_RI/ERI/UK_Ireland/Papers%20 and%20Written%20Consultation/20061106_Paper_IG2-7%20Analysis%20of%20interconnector%20 ﬂows_FI.doc

136

Competitive Electricity Markets

allocation of cross-border electricity interconnector capacity (i.e., congestion management), harmonization of transmission tariffs, and compensation to be paid to each TSO for hosting transmission ﬂows originating in other TSOs’ areas. It is clear that there are real limits on how much progress can be made in building on the regulation in moving quickly to the single market. Recent years have seen a signiﬁcant change of direction, with a new regional focus and a re-emphasis on structural and commercial rather than technical aspects of market access. The new regional initiatives under the auspices of the pan-European regulatory groups of CEER but especially Ergeg are endeavoring to prompt localized change as a catalyst to wider market integration. The development of regional markets as an interim step toward the establishment of a single European energy market is a sensible and probably the only viable way forward, and the potential for progress is clearly there if the process is supported politically by national governments. However, the scope for electricity transfers between member states remain constrained at less than 15% of consumption, and at the end of 2005 many interconnections remained managed by administrative rules without “market economic bases.”85 Furthermore, there is little early prospect that major new interconnection will change this context. Functional unbundling remains a signiﬁcant problem especially in the newer accessed states, and the commission is continuing to face sustained resistance to addressing remaining barriers in this area. It has concluded that deeper powers are required by national regulators to address this nexus of issues, but a further directive will be necessary to provide an appropriate legal basis. Against this background, producing a strategic vision paper in 2008 and a route map to a single market, which is what Ergeg has said it intends to do, looks extraordinarily ambitious. It is relatively early days, but the Western regional initiative has begun to make progress by focusing on priority actions in areas that presently distort interaction between markets. It is hard to determine whether other regions enjoy similar levels of commitment as there is limited public information as yet to assess these. This exercise has highlighted the types of issues that arise in considering development of a regional electricity market and the main areas of divergence that presently exist. How those issues will need to be addressed for a particular regional market, including fundamental questions of market design, is dependent on a number of technical, political, geographical, economic, and regulatory factors that are speciﬁc to that collection of national markets, and it is too early to say how these factors might impact across regions. However, it is possible to identify the key obstacles to the establishment and operation of a regional market and so provide a checklist of priority actions for further consideration in speciﬁc instances. It remains to be seen how and to what degree these priorities are effectively addressed by the ERI, and whether the process will be more successful than those initiatives that preceded it. As for the TSOs, they remain collectively an active participant in the process and the events of November 2006 have served to highlight the need for continuing convergence between physical operators. However, the need to join technical and trading solutions means that the national regulators are now in the driving seat. Further, the Ergeg process is beginning to achieve much wider participation by other stakeholders, including the PXs and representatives of the trader community, which will be essential if enduring solutions are to be identiﬁed.

EBL

85

Glachant, J-M. (2005). Implementing the European internal energy market in 2005–09, available at www.grjm.net.

Achieving Electricity Market Integration in Europe

137

What is informative are the areas of focus and the priority actions needed to increase the prospect of regional market convergence based on tackling congestion, intraday market access, and transparency, which should be of note to reformers not just in Europe but outside it. That said, these actions remain modest in scope and do not as yet embrace delivery of balancing arrangements that are transparent, simple, and robust to trading between markets. Further progress toward a single market design remains a long way off, and work in progress suggests that real progress toward this may still be a decade away, and that for the foreseeable future competition is likely, again to quote Glachant, to remain “fractured in national or local blocks.”

EBL

EBL

This page intentionally left blank

Part II Market Performance, Monitoring and Demand Participation

EBL

EBL

This page intentionally left blank

Chapter 4 Transmission Markets, Congestion Management, and Investment HARRY SINGH Federal Energy Regulatory Commission1 , Washington, DC, USA

Summary Nearly a decade has passed since the introduction of transmission open access in the United States under Order 888. During this period the electric utility industry has witnessed signiﬁcant changes in the transmission business with the development of organized markets operated by Regional Transmission Organizations (RTOs), the continued reﬁnement of the open access rules outside the organized markets, and the development of new transmission companies, such as merchant transmission projects and Independent Transmission Companies (ITCs). Transmission investment has generally lagged generation investment, putting greater emphasis in the United States on the role of price-based congestion management in market design than has occurred elsewhere. However, there is renewed focus on transmission with several initiatives to encourage investment and improve transmission markets. The August 2003 Northeast Blackout also drew attention to transmission operations and served as a catalyst for the development of enforceable reliability standards. This chapter reviews various approaches that have been implemented for transmission organizations in the United States using data from different markets. The chapter also provides an overview of signiﬁcant regulatory developments related to the Energy Policy Act of 2005.

EBL

4.1. Introduction The development of wholesale electric power markets has been driven largely by the goal of enabling competition in generation. However, connecting generation and load requires access to transmission. The transmission system consists of high voltage transmission lines, transformers, phase shifters, and other devices that enable the transfer of electric energy from generators to loads. This chapter explores the development of approaches to create open access to the transmission system and compares these approaches in terms of 1

The opinions expressed here are those of the author and do not reﬂect the view of the Federal Energy Regulatory Commission or any other organization that the author may have been afﬁliated with.

141

142

Competitive Electricity Markets

managing congestion and creating incentives for transmission investment. It also discusses the evolution of transmission organizations in the United States. This section provides background and introduces some major themes. A key development in creating transmission open access in the United States took place over a decade ago when Orders 888 and 889 were introduced by FERC in 1996, marking the ﬁrst attempt to standardize procedures for obtaining transmission service (FERC, 1996 and FERC, 1996a). Since that time, signiﬁcant changes have occurred in the structure of the transmission business that include the development of Independent System Operators (ISOs) or Regional Transmission Organizations (RTOs), the continued reﬁnement of the open access rules outside such organized markets, and the introduction of new transmission companies, such as merchant transmission projects and the formation of Independent Transmission Companies (ITCs). The separation of ownership from control of transmission assets in RTOs and ITCs has evolved differently in the United States as compared to other countries or regions where asset ownership and control are more typically combined within the same entity.2 Although pros and cons have been cited for both types of structures, the evolution of each was largely driven by the different starting points and differences in regulatory approaches as well as the intention of the policymakers (e.g., Chandley and Hogan, 2002; Henney and Russel, 2002; Ruff, 2002). A more detailed discussion of European Transmission System Operators (TSOs) can be found in Chapter 3. Section 4.2 of this chapter discusses the evolution of US transmission organizations and structures. In the US context, the evolution of ISOs and RTOs is closely linked to the creation of competition in generation. A useful analogy compares ISOs to air-trafﬁc controllers as entities in charge of ensuring safe and secure operation of the power grid (or airtrafﬁc) and impartially administering the rules of access. Although ISOs and RTOs are known as transmission system operators, they are just as much operators in the generation market.3 Given that vertically integrated utilities in the United States owned and operated both generation and transmission, there were three main options available for ensuring independence in transmission operation. The ﬁrst was to transfer the operational control of transmission assets to an independent entity such as an ISO or RTO; the second was to maintain control over transmission but offer transmission access and rates subject to the open access requirements; the third was to divest transmission assets to an independent company. As of this writing, approximately two-thirds of the US system has taken the ﬁrst approach, although the designs of ISOs and RTOs remain works in progress (see Fig. 4.1).4 The remainder of the system continues to take the second approach, which is also being continually reﬁned (e.g., FERC, 2007). While there are ITCs in the United States, they remain embedded in one or the other of these market models. Unlike the UK, no ITC in the United States undertakes congestion management under performance-based regulation. Sections 4.3, 4.3, and 4.5 of the chapter discuss transmission market design issues, such as congestion management, congestion metrics, and transmission rights, both within and outside the organized markets. The nature of RTO markets has evolved over time

EBL

2

Examples of transmission organizations that combine ownership and operations include the National Grid Company in the UK, Statnett in Norway, Red Eléctrica in Spain, and Transpower in New Zealand among others. 3 One ISO, the Independent Electricity System Operator (IESO) in Ontario was initially known as the Independent Electricity Market Operator (IMO) until it changed names in December 2004. 4 Roughly 60 per cent of the summer peak demand in the United States corresponds to regions with RTOs.

Transmission Markets, Congestion Management, and Investment

143

ISO-NE

Northwest NYISO

MISO

Peak load (MW)

PJM

CAISO

160,000 140,000 120,000

Southwest

SPP

100,000 80,000

Southeast

60,000 40,000 20,000

ERCOT

0 ISO-NE NYISO PJM

MISO

SPP ERCOT CAISO

Fig. 4.1. RTO and non-RTO regions. The northwest, southwest, and southeast are non-RTO regions. The projected peak demand for all regions in 2007 is 757 812 MW. RTO record peak loads in 2006. Sources: FERC, and NERC.

but appears to have settled on certain common elements. One such element is the use of auction-based markets for determining generation commitment and dispatch patterns (discussed in further detail in Chapter 5). These markets use a framework for congestion management that is known as Locational Marginal Pricing (LMP) or nodal pricing (the terms will be used interchangeably in this chapter). Along with LMP came allocations and markets for tradable Financial Transmission Rights (FTRs) designed to insulate transmission customers from the price risks created by congestion in an LMP market. Considerable debate ensued in determining the structure of LMP and FTR markets. With a few differences, these issues have largely been settled. For example, in the midand late 1990s, there was substantial concern in the nascent merchant and ﬁnancial power trading sector that the granularity of prices (i.e., prices set at each individual node instead of by zone) in LMP markets would have a signiﬁcant adverse impact on market liquidity. However, that did not prove to be the case. An example is the sustained liquidity of ﬁnancial trading at the PJM–West hub that aggregates individual LMPs at more than a hundred nodes. This experience built conﬁdence in LMP. In contrast, markets that started with zonal pricing (e.g., California and Texas) have transitioned to or are now gradually moving to a nodal approach. However, the FTR model has not been without its problems. For example, FTR prices have been quite volatile every year, reﬂecting volatility in congestion costs. This makes valuation difﬁcult. Also, multi-year FTRs were difﬁcult to design. Hence, markets for FTRs have seen the evolution of several new products. Section 4.3 discusses approaches for transmission markets while Section 4.5 describes FTR markets in particular. An important feature of the US markets is that the design of ISO and RTO markets varies by region, with major elements summarized in Fig. 4.1. The markets in the northeast, midAtlantic, and to some extent the Midwest are generally considered the most developed organized markets with full implementation of LMP.5 Other regions such as California

EBL

5 All these markets incorporate LMP implementations with day-ahead energy markets and FTRs. However, a majority of the schedules in the MISO day-ahead markets are still based on the commitment and schedule developed in utility control areas. Consequently, MISO is still moving toward implementing some market elements such as an Ancillary Services market.

144

Competitive Electricity Markets

and Texas incorporate varying elements of LMP markets and are in transition, while the Southwest Power Pool (SPP) operates a transmission market under Order 888 paradigm along with an auction-based real-time market.6 As noted, the rest of the country is served under open-access tariffs based on Order 888. PJM was the ﬁrst US market to implement an LMP-based day-ahead market on April 1, 1998 – one year after it started operating a regional bid-based market. NYISO started operation of an LMP-based market on December 1, 1999 followed by ISO-NE’s implementation of LMP on March 1, 2003. MISO implemented an LMP market on April 1, 2005. California is scheduled to implement an LMP market in 2008. Both PJM and ISO-NE initially started their markets without LMP but transitioned to an LMP approach after their initial experiences with congestion management. Table 4.1 lists key features and milestones in various RTO markets. While the United States has experimented with several kinds of transmission market designs and various regulatory approaches, a major indicator of policy and regulatory success is whether transmission investment has been seen as supportive of market development and improved reliability. These questions are explored in Section 4.6. Certainly the separation of generation from transmission that had historically been linked together in integrated resource planning has introduced new challenges to transmission planning. The further separation of ownership and operations in transmission and the interests of multiple entities have not made transmission planning and investment decisions any simpler. At the same time, the expansion of RTO footprints can make regional transmission planning easier by increased coordination across a wider geographical area. Transmission planning by RTOs is only now reaching a level of maturity that can be used to assess any real improvements. A key metric in transmission expansion decisions is the level of congestion on the grid. Increased price transparency in organized markets has made it easier to measure congestion. However, as discussed in Section 4.4, while several different congestion metrics have been proposed, not all metrics can be directly used as a measure on which to base investment decisions. The Northeast Blackout of August 2003 served as a catalyst for signiﬁcant action on the reliability front, such as the development of mandatory reliability standards and the creation of an Electric Reliability Organization (ERO). These developments followed the collapse of several merchant generation companies7 and also drew attention to transmission investment that had lagged behind generation investment for some time. Section 4.6 explains that although some merchant transmission projects have been implemented and several others have been proposed, they are unlikely to become the norm given the economies of scale that still exist in the transmission sector. Consequently, incentivebased regulated rates for transmission investments have received attention. Other recent actions such as a federal backstop authority in siting transmission projects and the repeal of the Public Utility Holding Company Act (PUHCA) are also likely to play a positive role in encouraging transmission investment.

EBL

6 Markets such as SPP have sometimes been characterized as “day 1” RTO markets as they largely operate under the Order 888 construct and are perceived as simpler and less expensive to implement. 7 Between January 2001 and January 2003, more than $130 billion was lost in equity value for just eight merchant generators including AES, Aquila, Calpine, Mirant, Dynegy, El Paso, Reliant, and Williams. Several merchant generation companies including Calpine, Mirant, NRG and National Energy & Gas Transmission ﬁled for Chapter 11 bankruptcy protection.

Transmission Markets, Congestion Management, and Investment

145

Table 4.1. RTO Markets RTO

Start Date

Markets

ISO-New England(ISO-NE)

1971 – New England Power Pool formed 1997 – ISO-NE formed 1999 – Initial market operation 2003 – LMP-based markets

LMP-based day-ahead and real-time markets with 8 load zones and 1 trading hub. Installed capacity market under transition to Forward Capacity Market (FCM), forward reserves market, regulation market, and FTR market.

New York ISO (NYISO)

1966 – New York Power Pool formed 1999 – LMP market started 2005 – Major software updated

LMP-based day-ahead and real-time markets with 11 zones. Zone J (New York City) and Zone K (Long Island) are major load pockets. Locational Installed Capacity market, ancillary services market, and FTR market.

PJM

1927 – PJM formed 1997 – PJM ISO formed 1998 – LMP market started 2004–05 – Major expansions

LMP-based day-ahead and real-time markets. Ancillary services market, Installed Capacity market under redesign, FTR market. Active trading hub at PJM–West.

Midwest ISO (MISO)

1998 – MISO formed 2002 – Order 888-based markets 2005 – LMP market started

LMP-based day-ahead and real-time markets. Ancillary services market, FTR market. Major trading hubs are Cinergy and northern Illinois.

Southwest Power Pool (SPP)

1968 1998 2004 2007

Order 888-based transmission markets and auction-based real-time market.

Electric Reliability Council of Texas (ERCOT)

1996 – Initial creation 2001 – Market operation 2009 – Market redesign using LMP

Real-time balancing and ancillary services markets, zonal congestion management with 4 zones, and FTRs.

California ISO (CAISO)

1998 – CAISO formed 2008 – Market redesign based on LMP

Markets for real-time energy, day-ahead transmission, ancillary services, and FTRs.

EBL

– – – –

SPP formed Tariff administration RTO status Real-time LMP market

Source: FERC, 2007.

4.2. Evolution of Transmission Organizations and Structures Vertically integrated utilities have historically operated a large portion of the transmission system across the United States.8 Although analogies are sometimes made between the transmission system and the interstate highway system, historically transmission planning 8 There are almost 450 different transmission owners and approximately 135 control areas or balancing authorities across the country. A complete list of balancing authorities can be found at http://www.nerc.com/∼org/

Competitive Electricity Markets

146

Table 4.2. Major milestones in the US power industry Date

Event

Comments

1935

Federal Power Act

Enacted by US Congress and basis for FERC jurisdiction over interstate transmission and wholesale power rates

1965

Northeast Blackout

Provided impetus for major reliability initiatives including formation of NERC

1968

North American Electric Reliability Council (NERC) formed

Non-proﬁt organization formed by utility industry to promote reliability and adequacy of bulk power supply. In 2006, NERC was certiﬁed as the Electric Reliability Organization (ERO)

1978

Public Utility Regulatory Policy Act (PURPA)

Required utilities to purchase power from Qualifying Facilities (QFs) at avoided costs.

1992

Energy Policy Act of 1992 (EPACT 1992)

Created category of non-utility generators and required transmission open access

1996

FERC Orders 888 and 889

Foundation of transmission open access

2000

FERC Order 2000 and creation of RTOs

Proposal for voluntary formation of RTOs

2000–01

California power crisis

Lack of long-term contracts in face of high spot prices led to bankruptcy of PG & E, setback for expansion of competitive markets. Markets stable since then albeit new market design proposed for 2008.

2001

Enron Bankruptcy

Further setback for competitive markets, credit downgrades that followed impacted long-term contracting and trading, banks and ﬁnancial institutions stepped in later to ﬁll the void left by Enron.

2002

FERC Standard Market Design (SMD)

Intended to avoid repeat of California experience and help address seams issues by standardizing best practices in RTO market design. However, efforts to expand RTOs into new areas led to resistance and SMD was withdrawn in 2005.

2003

August 14 Northeast Blackout

Largest power outage in North America with more than 61 800 MW of lost load affecting an estimated 50 million people.

2005

Energy Policy Act of 2005 (EPACT 2005)

Several initiatives to address reliability, transmission infrastructure, and market manipulation.

2007

Mandatory Reliability Standards

Mandatory standards took effect on June 18, 2007. NERC certiﬁed as Electric Reliability Organization (ERO) in 2006.

EBL

was done primarily within the footprints of large vertically integrated utilities with the goal of serving local load. The evolution of the transmission system from isolated standalone systems to regional systems and ultimately to interconnections took place over many decades (Casazza, 1993). This paradigm changed signiﬁcantly starting in 1978 with the Public Utilities Regulatory Policies Act (PURPA), which was followed by the Energy Policy Act of 1992 (EPACT 1992), and ﬁnally by Orders 888 and 889 in 1996. Seeking to promote the use of renewable energy, PURPA opened up the market for independent power producers by requiring

Transmission Markets, Congestion Management, and Investment

147

utilities to purchase power from qualifying facilities (QFs) at avoided costs. EPACT 1992 expanded wholesale power sales to non-utility generators called Exempt Wholesale Generators (EWGs). However, there was little standardization in the procedures for obtaining transmission access until FERC Order 888 (FERC 1996, 1996a) that required utilities to provide open access to the use of their transmission system on the same terms and rates as available to the utility’s native load. Order 888 also required utilities to functionally unbundle their rates for wholesale generation and transmission service. Various ancillary services were unbundled and included as a component of required transmission service. Although this had signiﬁcant impact on the number and nature of power transactions, there was little structural change in the transmission business itself. Subsequent regulatory initiatives advanced the move toward formation of transmission companies. An important question in this movement was whether ownership and control of transmission needed to stay together. The Federal Power Act (FPA) established by Congress in 1935 has been administered by FERC to ensure that wholesale power rates are “just and reasonable.” The introduction of competition has seen changes in the FPA most recently with the Energy Policy Act of 2005 that enhanced FERC’s authority in several areas such as enforcement of mandatory reliability standards and created a backstop siting authority for transmission. However, the “just and reasonable” standard continues to remain a major component of the FPA despite the transition from cost-of-service ratemaking to competitive markets. Following several years of experience under Order 888, FERC perceived that the continued balkanization of the United States grid under different transmission operators along with concerns about interpretation or possibly manipulation of open access rules was creating market inefﬁciency. Hence, it established a new requirement, that of Regional Transmission Organizations (FERC, 1999). The intention was to ensure that all US utilities under FERC’s jurisdiction would join large regional RTOs. However, in some regions of the country there was resistance to the RTO concept, and this resistance grew during the California power crisis in 2000–01. The role of regulators then changed quickly and gained increased prominence following the California power crisis. The signiﬁcance of good risk management and well-designed markets became increasingly evident. The evolution of markets in different regions with different rules also created “seams issues” between different transmission organizations. In 2002, FERC proposed to establish a Standard Market Design (SMD) for all utilities under its jurisdiction in part to address the seams issue and in part to avoid a repeat of the California crisis that was attributed to ﬂawed market rules.9 Perhaps a more signiﬁcant component of SMD was the goal of advancing RTO

EBL

9 Although the California crisis is often linked to ﬂawed market rules, one of its major factors was the absence of any long-term contracts. In contrast, other market shortcomings such as the lack of a centralized day-ahead market and LMP played only a secondary role. In fact, some of these shortcomings continue to this day in equally tight supply conditions but without any major disruptions due largely to the presence of long-term contracts and improved market oversight and mitigation. It is also noteworthy that the California and Texas wholesale markets shared similar market designs that lacked centralized day-ahead markets and LMP. The characterization of one as a failure and the other as a success was more a consequence of the different fundamentals and levels of long-term contracting in each market. Further details on these markets can be found in Sweeney (2006) and Adib and Zarnikau (2006). While California enjoys a relatively high level of forward contracting today, other states like Maryland and Illinois that are participants in the generally well-regarded PJM market are facing retail rate-shocks due to expiration of buy-back contracts at a time of high wholesale prices, expiration of retail caps, and retail rate design. Even if this is reminiscent of the California experience, the impacts are likely to be less severe given the maturity of the PJM market.

Competitive Electricity Markets

148

participation. However, as with the RTO initiative, regional dissatisfaction with what was perceived to be an overly prescriptive federal approach created strong resistance in many states to adopting SMD. The resistance to further expansion of RTO markets was also aided by the California experience and the Enron bankruptcy. The SMD proposal was eventually withdrawn by FERC in 2005, but it remained inﬂuential in promoting the LMP and FTR market design, as discussed in Sections 4.3–4.5. Table 4.2 lists some major milestones in the US power industry.

4.2.1. Transcos, ITCs, and ICTs There are various business structures possible for a transmission company. A transmission company could be a for-proﬁt or a non-proﬁt organization; it could combine ownership and operation of assets or separate them into different entities. Organizationally, RTOs are independent non-proﬁt entities with operational control but no ownership of assets. This is in stark contrast to successful for-proﬁt transmission operators in Europe that both own and operate transmission assets. The non-proﬁt aspect of RTOs was not explicitly required by FERC Order 2000 that initiated their formation; rather it was a consequence of the debate that followed. The characteristics required of an RTO included independence, regional scope, operational control, and regional reliability (see Box 4.1 for characteristics and functions of RTOs as speciﬁed in FERC Order 2000). For-proﬁt transmission companies that combined ownership and operation, referred to as Transcos, were proposed as potential RTOs but never implemented in the United States. There have been recent successes with ITCs that consolidate transmission ownership of transmission assets but still operate under RTOs.

EBL

Box 4.1 Main characteristics and functions of RTOs RTO Characteristics (FERC Order 2000) • • • •

Independence Scope and regional conﬁguration Operational authority Short-term reliability

RTO Functions (FERC Order 2000) • • • • • • • •

Tariff administration and design Congestion management Parallel path ﬂow Ancillary services Available Transmission Capability (ATC) Market monitoring Planning and expansion Inter- regional coordination

Transmission Markets, Congestion Management, and Investment

149

Table 4.3. Stand-alone transmission companies Transmission Company

Ownership

Operational Control

Established

Areas

ITC Transmission

Owned by ITC Holdings held by private equity investors. Formerly owned by DTE Energy, the parent company of Detroit Edison.

MISO (2700 miles of high-voltage transmission)

2003

Michigan, parts of Ontario, and other Midwestern states

METC (Michigan Electric Transmission Company)

Owned by ITC Holdings held by private equity investors. Past owners include Trans-Elect and Consumers Power

MISO (5400 miles of high-voltage transmission)

2002

Lower peninsula of Michigan

American Transmission Company ATC

Owned by member utilities, municipals, and cooperatives

MISO (9081 miles of high-voltage transmission)

2001

Eastern Wisconsin and parts of Michigan, Illinois, and Minnesota

Source: ATC, ITC Holdings 2006.

EBL

Table 4.3 lists the various stand-alone transmission companies in operation today. The ﬁrst independently owned and operated transmission company in the United States began operations in 2003 as the International Transmission Company (ITC Transmission) in southeastern Michigan. Since then ITC Transmission’s parent company, ITC Holdings, acquired additional transmission assets from an adjacent transmission company, Michigan Electric Transmission Company (METC) and recently announced plans to acquire transmission assets from Alliant Energy that would expand the company’s footprint into other Midwestern states including Iowa, Illinois, and Minnesota. The ﬁrst transmission-only utility in the country also started in the Midwest as American Transmission Company (ATC) in 2001. However, unlike ITC-Holdings which is owned by private equity ﬁrms Kohlberg Kravis Roberts (KKR) and Trimaran Fund Management, ATC is owned by the utilities, municipal electric companies, and electric cooperatives that contributed their transmission assets to ATC. Both companies are members of the Midwest ISO. There is also a recent proposal to form an ITC in Kansas (ITC Great Plains) within the SPP footprint. The deﬁnition of Transcos and ITCs has continued to evolve since these terms were ﬁrst used (Box 4.1). More recently, FERC has used the term Transco more broadly to deﬁne any stand-alone transmission company that sells transmission services at wholesale and/or on an unbundled retail basis. This deﬁnition includes independent transmission companies like International Transmission as well as those afﬁliated with other market participants like American Transmission Company (FERC, 2006a). The debate on which structure makes most sense has been inﬂuenced by legacy and independence considerations in the United States. In some instances, vertically integrated utilities elected to turn over operation of their transmission system to an RTO as it could address independence concerns without having to divest any generation assets. In other

150

Competitive Electricity Markets

instances, for-proﬁt Transcos were proposed on the grounds that they could provide more value and innovation by combining ownership and operation.10 However, such proposals met strong resistance from some stakeholders on their ability to be truly independent on policy decisions between transmission investment and generation or demand response. For example, generators in a load pocket might fear that excessive transmission upgrades may adversely impact their revenues. There were also questions on the market designs for short-term operations under a Transco – e.g., could it implement an LMP-based market like an RTO? Some of these concerns were addressed by ITCs that offered the advantages of horizontal consolidation of transmission assets but would turn over operations to an RTO to address independence concerns. Transco proponents argued that turning over operations to the RTO would dilute the value proposition of a Transco and make the transmission owner passive. However, a transmission owner operating under an RTO may not be any more passive than a generation owner operating under the RTO. Even though a generator receives dispatch instructions from the RTO, one seldom hears of passive generation owners. A more accurate characterization might be that the fewer functions a transmission owner performs, the more limited the options will be for designing incentive regulation proposals. Another more recent development in transmission organizations involves the creation of an Independent Coordinator of Transmission (ICT), by which a transmission owning utility arranges for an independent entity to coordinate the operation of its transmission system. The utility still remains the transmission owner and operator but certain limited functions such as granting or denying transmission service requests, calculating available ﬂowgate or available transfer capacity, processing generator interconnection requests, and operating the OASIS11 website are outsourced to the ICT. The ﬁrst example of an ICT involves a structure in which SPP will serve as the ICT for Entergy, a utility in the southeastern United States. In contrast with RTOs that operate daily and hourly markets, the Entergy–SPP ICT will operate a weekly procurement process for energy that is intended to allow independent generators to compete with the incumbent utility (Entergy). Other ICT examples involve Duke Energy, where MISO will act as the Independent Entity (IE), and MidAmerican, where TranServ will act as the Transmission Service Coordinator (TSC). Under a similar arrangement, SPP serves as the Independent Transmission Organization (ITO) for Louisville Gas and Electric and Kentucky Utilities. Figure 4.2 illustrates control and ownership of transmission under different organizational structures.

EBL

4.2.2. Incentives for RTOs Given the non-proﬁt structure of RTOs, the question of performance incentives can be important. Incentive regulation plans designed for a for-proﬁt transmission operator do not automatically lend themselves for application to a non-proﬁt entity. However, it is possible to deﬁne metrics for measuring the performance of a transmission operator that could be used in setting management compensation. Table 4.4 lists examples of potential metrics that may be considered (Singh, 2005). Increases in RTO costs have in some cases dampened enthusiasm to join RTOs and there have been recent efforts to reduce these costs. For example, the Grid Management 10

The proposal for the Alliance RTO in the Midwest is one example of a Transco that did not materialize. 11 OASIS or Open-Access Same-Time Information System represents the Internet-based system used to obtain transmission service and related information.

Transmission Markets, Congestion Management, and Investment

RTO

Integrated utility

Non-profit entity control of operations no asset ownership

Non-profit entity control of operations no asset ownership

151

Transco For-profit entity transmission owner control of operations

ITC

ICT

For-profit entity consolidates asset ownership limited operational role

Independent entity for-profit or non-profit Performs limited operational functions for utility

Fig. 4.2. Control and ownership under different transmission organizations. All the existing ITCs operate under the umbrella of an RTO that is responsible for system operations. However, ITCs in non-RTO regions may be more akin to a Transco that combines ownership and operations.

Table 4.4. RTO metrics Metrics for measuring RTO performance

EBL

Accuracy of load forecasts Compliance with reliability standards Accuracy of settlements Settlement cycle Frequency of price revisions Frequency of tariff violations

Frequency of tariff changes Magnitude of administrative costs Magnitude of uplifts charges Efﬁcient utilization of grid Accuracy of information Customer focus

Charge (GMC) at the CAISO has decreased after being among the highest of all RTO charges. Overall, the average administrative and operating charges across all RTOs was approximately $0.5/MWh in 2005 amounting to more than $1 billion in annual payments by market participants. Even though this is a signiﬁcant ﬁgure, RTO costs comprise less than one per cent of the overall cost for power and are still small compared to cost overruns in the earlier era of cost of service regulation. Most RTO markets involve commitment of resources to meet the RTO’s load forecast. If the RTO’s forecast is inaccurate, excessive costs may be incurred that are eventually allocated to load as uplift costs. Thus, accuracy in load forecasting would translate to increased efﬁciency in the operation of an RTO. Similarly, the performance of an RTO as a control area operator or balancing authority can be measured by frequency control metrics such as CPS1 and CPS2 deﬁned by NERC.12 PJM reported improved control performance 12

CPS1 and CPS2 (Control Performance Standards) are reliability standards that measure the performance of a control area with respect to controlling system frequency and area control error. Under the proposed mandatory reliability standards, CPS1 and CPS2 will be replaced by a new standard for Balance of Resources and Demand BAL-007-1.

152

Competitive Electricity Markets

after the introduction of LMP markets but the performance declined somewhat after PJM’s expansion into new regions. Another metric could be the level of demand response in the RTO. Price sensitive demand response can play an effective role in improving market efﬁciency and considerable cost savings during system peaks.13 The proper integration of demand response with pricing in RTO markets is a signiﬁcant issue that continues to develop. Further discussion of demand response can be found in Chapter 8. Most RTO markets involve a time lag between the sale/consumption of energy and actual ﬁnancial settlement due to the time it takes for actual meter reading. This can increase credit risk as well as the collateral requirements for market participants. Efforts have been made by some RTOs to shorten their settlement cycles to address these concerns although there is still not a uniform standard. The early operation of some RTO markets witnessed frequent revisions to published prices that introduced an added dimension of risk to transaction ﬁnality. Equally challenging is the continuous change in market rules. For example, the CAISO has had at least 73 amendments to its original tariff. Although market design changes often improve market functioning, frequent changes create regulatory uncertainty that also adds risk. Prices from RTO markets can provide a useful index for use in long-term contracts. Nowhere has this been as evident as in PJM where the volume of ﬁnancial swaps indexed to the PJM–West hub has increased dramatically. This reﬂects the conﬁdence of market participants in the published price. In cases where the published price does not reﬂect certain costs known as uplifts, its usefulness as an index can decrease. The debate on RTO responsiveness has also involved discussion on what form of governance makes the most sense. With the exception of the CAISO during its initial start-up, RTOs have been governed by independent boards. Independent boards have sometimes been criticized as not being completely in touch with the needs of market participants and beholden to RTO management. In contrast, boards that involve representation of market participants or stakeholders can become beholden to commercial interests of member companies. There has been discussion of hybrid board structures, advisory committees, and the need for access to board members to address some of these issues.14

EBL

4.2.3. Reliability In addition to the evolution of transmission business structures, another area of change deals with reliability standards. All transmission organizations must operate in compliance with standards established by NERC. Until recently, these reliability standards have largely been implemented on a voluntary basis through NERC’s member reliability councils. In 2006, the number of reliability councils was reduced from ten to eight when three councils merged to improve consistency of reliability criteria across the region. A new reliability council, called Reliability First, was formed from the merger of East Central Area Coordination Agreement (ECAR), Mid-Atlantic Area Council (MAAC), and MidAmerican Interconnected Network (MAIN). The hierarchy of entities on the reliability 13

On August 2, 2006, PJM experienced a record peak load of 144,796 MW when voluntary reductions in electricity consumption produced savings of $230 million. Total savings from demand response during a 1-week period exceeded $650 million. 14 Governance issues have been an area of focus in a recent strategic initiative by PJM to examine improvements in its markets http://www.pjm.com/documents/strategic-report.html

Transmission Markets, Congestion Management, and Investment

153

QUÉBEC INTERCONNECTION

NERC INTERCONNECTIONS

NPCC MRO

RFC WECC SPP SERC

FRCC WESTERN INTERCONNECTION ERCOT

EASTERN INTERCONNECTION

ERCOT INTERCONNECTION

Fig. 4.3. NERC Interconnections and Reliability Councils. Source: NERC.

EBL

front starts with the interconnections. NERC spans four interconnections in North America of which three are in the United States – the western and eastern interconnections and Texas. Within each interconnection are several regional councils, as shown in Fig. 4.3. There are also several reliability coordinators listed in Table 4.5 as well as multiple balancing authorities. In response to the August 14, 2003 Blackout in the Northeast, EPACT 2005 called for mandatory and enforceable Reliability Standards. In February 2006, FERC issued Order 672 that certiﬁed NERC, as the Electric Reliability Organization (ERO) (FERC, 2006b). The ERO is required to develop Reliability Standards, which are subject to FERC review and approval. Since then NERC has proposed 102 standards based largely on its existing practices for certiﬁcation with FERC. In October 2006, FERC proposed adoption of 83 of these standards as mandatory and enforceable standards (FERC, 2006c, 2007a). The mandatory standards took effect in June 2007.15 The remaining continue to apply on a voluntary basis. The standards span several different categories as shown in Table 4.6. The primary responsibility for enforcing reliability standards rests with the regional councils and the ERO although FERC has ultimate authority. This is in contrast to FERC’s more direct and active role in enforcement of market rules and preventing market manipulation. However, FERC’s enhanced authority under EPACT 2005 to levy civil penalties as high as $1 million/day for each violation of the Federal Power Act applies to both market and reliability rules. In addition to approving mandatory reliability standards, FERC also approved violation risk factors that classify the consequences 15

The standards were initially set to take effect on 4 June 2007 but the effective date was pushed back to 18 June 2007.

Competitive Electricity Markets

154

Table 4.5. NERC Regional Reliability Councils Reliability Council

Description

Electric Reliability Council of Texas (ERCOT)

Single Balancing Authority, covers Texas

Florida Reliability Coordinating Council (FRCC)

Eleven Balancing Authorities, covers Florida

Midwest Reliability Organization (MRO)

Multiple Balancing Authorities in eight US states and two Canadian provinces

Northeast Power Coordinating Council (NPCC)

Covers 20% of load in eastern Interconnection across northeastern US and central and eastern Canada

Reliability First Corporation (RFC)

Twelve Balancing Authorities across multiple US states.

SERC Reliability Corporation (SERC)

Covers sixteen southeastern and central states and contains 5 sub-regions – Southern, Entergy, TVA, VACAR, and Gateway

Southwest Power Pool (SPP)

Covers southwest quadrant, also operates as Reliability Coordinator and RTO, multiple Balancing Authorities

Western Electricity Coordinating Council (WECC)

Largest NERC region, covers entire Western Interconnection

EBL

Source: NERC.

Table 4.6. Categories for reliability standards Group

Description

BAL CIP COM EOP FAC INT IRO MOD PER PRC TOP TPL VAR

Resource and Demand Balancing Critical Infrastructure Protection Communications Emergency Preparedness and Operations Facilities Design, Connections, Maintenance, and Transfer Capabilities Interchange Scheduling and Coordination Interconnection Reliability Operations and Coordination Modeling, Data and Analysis Personnel Performance, Training, and Qualiﬁcations Protection and Control Transmission Operations Transmission Planning Voltage and Reactive Control

Source: NERC.

of non-compliance into low, medium, and high risk (FERC, 2007b).16 Violation risk factors together with the severity of the violation will be used by NERC to establish penalties for non-compliance. EPACT 2005 also gave FERC reliability authority over nonjurisdictional entities such as municipals, cooperatives, federal power marketing authori16

FERC proposed modiﬁcations to 28 of more than 700 violation risk factors proposed by NERC.

Transmission Markets, Congestion Management, and Investment

155

ties, and utilities in Texas that are normally not within FERC’s jurisdiction. The move to transition voluntary reliability rules into enforceable standards is also receiving attention in Europe.17 4.3. Approaches for Transmission Markets US transmission markets remain divided between the regions outside the organized markets, which will be called “Order 888-based markets” in this chapter, and those with centrally organized markets, most of which follow a design with LMP and FTRs. There are also hybrid markets emerging that share features of both. This section ﬁrst describes the former type and then the latter, followed by some additional details on LMP. More detail on FTRs follows in Section 4.5. 4.3.1. Order 888-based markets Transmission service can be deﬁned as the right of a customer to inject energy at one or more points on a transmission service provider’s electric grid (“points of receipt”) and withdraw energy at one or more points on the service provider’s electric grid (“points of delivery”). The ﬁrst attempt to standardize the rules for transmission service came with Order 888. Transmission access under Order 888 is based on a ﬁrst-come ﬁrst-served approach to physical access that presents little price risk but can involve risks of curtailment via Transmission Loading Relief (TLR) procedures.18 The key constructs of Order 888 include the availability of point-to-point and network service, varying levels of ﬁrmness (e.g., ﬁrm or non-ﬁrm) and different term-lengths of service (e.g., hourly, monthly, or annual). Under point-to-point service, a transmission user must designate speciﬁc points of receipt and delivery. Network service does not require speciﬁc points of receipt and delivery but rather the speciﬁcation of network loads and resources. Figure 4.4 depicts the choices for transmission service under Order 888. In addition to the risk of curtailment under TLRs, questions have been raised about the transparency of Available Transfer Capability (ATC)19 calculations that can be critical in obtaining physical access to the grid and the use of a contract-path paradigm. The development of organized markets operated by RTOs described next can be viewed as an

EBL

17 The Union for the Co-ordination of Transmission of Electricity (UCTE), an association of Transmission System Operators (TSOs) in 23 European countries, adopted a multilateral agreement in June 2005 to make reliability standards binding and enforceable with inter-TSO penalties for damages up to E5 million. The second phase of this effort involves updating and modifying existing standards similar to what is planned in North America. 18 TLRs have been characterized as an un-scheduling procedure that was developed to address inaccuracies in scheduling of transmission service. TLRs have been criticized as not always being efﬁcient (e.g., Rajaraman and Alvarado, 1998). 19 ATC = TTC – ETC – CBM – TRM where TTC is Total Transfer Capability, ETC denotes Existing Transmission Commitments including serving native load, CBM denotes Capacity Beneﬁt Margin, and TRM denotes Transmission Reliability Margin. In addition to the lack of transparency in models deﬁning ATC calculations, assumptions on the various inputs can also inﬂuence the calculation. Overly conservative assumptions on ETC, CBM, and TRM can reduce ATC unnecessarily. There are at least three different approaches currently in use to calculate ATC. These include network ATC, network AFC (Available Flowgate Capacity), and contract path approaches. The contract path approach has been largely used in the Western Interconnection.

Competitive Electricity Markets

156

Network integration service

Point to point (PTP)

Non-firm

Monthly weekly daily hourly

Firm

Short term

Long term (LT)

Monthly weekly daily

Annual or longer

Fig. 4.4. Types of transmission service under Order 888.

attempt to address some of the concerns about Order 888. However, after recognizing that the further expansion of RTO markets may not occur, attention was given to less radical improvements to Order 888. FERC recently proposed reforms under Order 890 that retain the underlying ﬁrst-come ﬁrst-served construct of Order 888 but attempt to address some of the concerns that have been raised by market participants (FERC, 2007). Major areas of reform in Order 890 include a requirement for increased transparency for ATC calculations, elimination of imbalance penalties, introduction of a new conditionalﬁrm service that would avoid denying transmission requests that are infeasible for only a short period, further clarity in re-dispatch obligations, and a reform of the rollover rights policy.20 A new product in the form of hourly ﬁrm service was initially considered but not adopted in the ﬁnal version of Order 890. There are also requirements for increased transparency in intra- and inter-regional transmission planning. Criticisms of Order 888 being based on a contract-path approach have more to do with the geographic scope of ATC calculations rather than the underlying construct itself. In this sense, an Order 888based market operated over a large region by an independent entity (e.g., the SPP RTO) can also serve as an example of Order 888 reform as it addresses both the contract path and the independence concerns.

EBL

4.3.2. Transmission access in organized markets Transmission access in most organized RTO markets (with the exception of SPP, as discussed later) is a three-step process in roughly the following sequence: First, an application for transmission service and payment of an access charge (which is used to compensate rate-based transmission assets in the RTO footprint); second, allocation and/or auction of ﬁnancial transmission rights (FTRs), typically done months or weeks prior to the dispatch 20

Further details on Order 890 can be found at http://ferc.gov/industries/electric/indus-act/ oatt-reform.asp

Transmission Markets, Congestion Management, and Investment

157

day; and third, actual transmission usage, which is subject to marginal congestion and loss charges derived from the day-ahead and real-time energy market LMPs. FTRs are then used to hedge the congestion component of the LMPs. The transmission markets are also open to ﬁnancial (“virtual”) trading by non-transmission users, whether seeking to purchase FTRs or to take positions in the energy markets. In a sense, the LMP markets are simultaneous auctions of transmission capacity under which offers and bids for energy establish generation and load schedules subject to transmission constraints. Physical access to the grid is ensured under this approach but comes with a higher price risk that can be managed through FTRs. The allocation of transmission capacity in the form of FTRs is conceptually similar to the allocation of the same capacity as physical transmission rights under Order 888, albeit over the larger footprint of an RTO. Currently, FTRs are limited to annual terms but there are efforts underway to offer long-term FTRs that span multiple years. A major difference in the early implementation of RTO markets was the granularity of locational prices. Some markets such as California and Texas initially opted for market designs with prices aggregated across zones. Experiences with uplift costs and the subsequent redeﬁnition of zones have caused these markets to move closer to a full nodal pricing framework similar to the markets in the Northeast and Midwest. In principle, an RTO could follow the Order 888 approach to allocating physical transmission capacity over the territory of its members. This would overcome some of the criticisms of Order 888 markets such as contract-path limitations and independence. Some RTOs, such as SPP and MISO during their initial start-up, have followed such approaches. Figure 4.5 summarizes the different types of RTO and non-RTO markets. Non-RTO markets can sometimes be separated into those that border RTOs and those that do not. There has been recent discussion of RTO beneﬁts accruing to neighboring non-members at a

EBL

Auction based markets 888 markets 888 regions that are RTO neighbors

888 regions that are not RTO neighbors

RTOs-day-ahead & real-time markets

PJM NYISO ISO-NE MISO

RTOs-realtime markets

CAISO ERCOT

RTOs 888 rights

LG&E BPA Entergy TVA MISO

Southern FPL Avista APS

SPP

Fig. 4.5. Different transmission structures in the United States. At one end of the spectrum are RTOs that operate auction-based day-ahead and real-time markets, followed by RTOs that operate auctionbased real-time markets. At the other end are vertically integrated utilities that provide transmission access under Order 888. Somewhere in the middle lie RTOs such as SPP that provide transmission access under Order 888 but also operate auction-based real-time markets. MISO contains both a market footprint with auction-based markets and a larger reliability footprint that also contains Order 888-based markets.

158

Competitive Electricity Markets

lower cost than what is borne by RTO members.21 For example, a utility neighboring an RTO can buy or sell power from the RTO market paying transaction costs only on its net sale or purchase. In contrast, utilities within the RTO pay such costs on their gross demand. This is only one example of the many seams issues between transmission organizations.

4.3.3. Features of locational marginal pricing As noted, two major elements of organized power markets are the use of single price auctions and LMP-based congestion management. In the absence of congestion (and ignoring the impact of losses), power prices in all locations are the same and set by the most expensive resource selected to serve load at a given point in time. This aspect of organized markets is discussed further in Chapter 5. The presence of congestion causes generation to be re-dispatched and power prices to vary by location. In general, prices in LMP-based wholesale power markets vary by location and time and reﬂect the incremental cost of meeting demand at any location at any point in time. In some implementations such as in New York and the one proposed in California, loads see prices that are averaged across zones. The emphasis on LMPbased markets in the United States speaks to the signiﬁcance of congestion management in overall market design. A principal beneﬁt of an LMP market is that it provides a transparent means to price and allocate the use of the transmission system. This provides operational beneﬁts as well as visible price signals for new investment and demand response. One operational beneﬁt is fewer instances of administrative curtailment procedures such as TLRs. In terms of impacts on investment, markets without LMP that socialize congestion costs may not create the best incentives for siting new generation in the right locations. High levels of intra-zonal congestion in southern California in 2004 that also raised reliability issues have at least in part been attributed to poor siting decisions by new generators (California ISO, 2004a). While some intra-zonal congestion was subsequently alleviated by transmission upgrades, it could recur unless the underlying problem of price signals is addressed. LMP markets have been accepted as the preferred approach in most organized power markets. The alternative to LMP is the adoption of a single price across the market with transmission congestion cost allocated as a separate uplift charge that is not reﬂected in the market price. A major drawback of such approaches is that uplift costs can be unpredictable and are difﬁcult to hedge through long-term contracts. Markets such as California and Texas tried zonal approaches but faced problems with congestion causing signiﬁcant uplift costs and the need to redeﬁne zonal boundaries. An alternative approach was implemented initially by ISO-NE, where the entire market was treated as a single zone and all out-of-merit generation was paid through an uplift. Additionally, non-LMP approaches that do not directly assign the cost of congestion to those who contribute to it can lead to gaming, as was observed during the California power crisis. One such game involved the submission of infeasible schedules that caused

EBL

21

See Louisville Gas and Electric Company, et al., 114 FERC ¶ 61,282 (2006) at p. 64–65 approving the withdrawal of Louisville Gas and Electric Company (LG&E) and Kentucky Utilities (KU) from the Midwest ISO (MISO). MISO had argued that the request should be denied because the utilities “propose to beneﬁt from the ongoing existence of the Midwest ISO, while paying for the services offered by the Midwest ISO only on a selective, as-needed basis.”

Transmission Markets, Congestion Management, and Investment

159

or exacerbated intra-zonal congestion followed by payments for decrementing output to help relieve the same congestion. This so-called “DEC” game was made possible by a combination of market power and the speciﬁcs of market rules that were in place at the time.22 Such problems have not been observed in LMP markets. LMP implementations in several organized markets also incorporate marginal loss pricing as a component of LMPs. This means that LMPs can vary across locations even in the absence of congestion. NYISO, ISO-NE, MISO, PJM, and CAISO all either already incorporate marginal losses in LMPs or are moving in that direction. Transmission losses can be approximated as a quadratic function of the power injected into the transmission system (Wood and Wollenberg, 1996). Consequently, the inclusion of marginal losses in LMPs creates surplus revenue that is approximately twice the average cost of losses. Unlike congestion which only occurs when ﬂows on lines reach binding limits, losses are present at all hours of operation. Consequently, the loss surplus can be considerable and even comparable to the congestion surplus. Unlike congestion rents that are used to fund FTRs, the surplus from marginal losses is currently refunded to market participants using different approaches. Loss hedging rights have been proposed but not implemented (e.g., Harvey and Hogan, 2002; Rudkevich et al., 2005). The basics of LMP can be illustrated using a simple example where three locations or nodes are connected by identical transmission lines (Fig. 4.6) (additional examples are found in Chapter 5.). Cheap generation is available at node 1 at $20/MWh. More expensive generation is available at node 2 at $30/MWh. Each generator has a capacity of 400 MW. A ﬁxed load of 450 MW is at node 3. In the absence of any transmission constraints, the

EBL (1/3) x 180 = 60 MW

Unit A 270 MW

Unit B

1

$20/MWh 180 + 60 = 240 MW

2 (1/3) x 270 = 90 MW (2/3) x 270 = 180 MW

180 MW $30/MWh

(2/3) x 180 = 120 MW

3 Optimal dispatch: Unit A = 270 MW Unit B = 180 MW All lines have equal impedance No losses

450 MW

Prices: Node 1: $20/MWh Node 2: $30/MWh Node 3: $40/MWh

Fig. 4.6. LMP calculation in a three-node network.

22

See, e.g., 103 FERC ¶ 61, 265 Order on proposed tariff amendment 50, May 30, 2003 and Opinion of the ISO Department of Market Analysis on the need for alternatives to economic bids for managing intra-zonal congestion in the absence of competitive conditions for re-dispatch, December 20, 1999.

160

Competitive Electricity Markets

cheaper generator at node 1 is dispatched at its maximum capacity of 400 MW and the expensive generator at node 2 supplies the remaining 50 MW. The LMP at any location is deﬁned as the cost of meeting an additional MW of demand at that node. Ignoring transmission constraints, the next MW of supply comes from the expensive generator and consequently the price is $30/MWh at all locations. Next, assume a capacity limit of 240 MW applies from node 1 to node 3. In this case the cheaper generator can no longer operate at its maximum output as that would cause the power-ﬂow in line 1–3 to exceed its limit. The optimal dispatch involves dispatching the cheaper generator at 270 MW and the expensive generator at 180 MW. The output of each generator splits up between the two parallel paths in inverse proportion to the distance involved. Thus, two-thirds of the power from the generator at node 1 ﬂows directly from node 1 to node 3, while the remaining one-third ﬂows through node 2. The LMP at the three nodes under the constrained case is $20, $30, and $40/MWh, respectively. This is easy to interpret for each of the generator nodes as a megawatt of demand would be supplied by local generation. For node 3, an increase in demand involves increasing the output at the expensive generator by 2 MW and reducing the output at the cheaper generator by 1 MW giving the $40/MWh result. This example also illustrates the fact that LMPs can at times exceed the bid caps applied to generators. The overall impact of congestion is to increase LMPs at the load and decrease it at generator node 1. Thus, any loads located in generation pockets or export regions can actually see lower prices due to congestion while loads in load pockets or import regions see increased prices. The alternative to LMP would be to socialize the re-dispatch costs. In this example, Unit A’s output is reduced by 400 − 270 = 130 MW while Unit B’s output is increased by 180 − 50 = 130 MW. The net change in cost is characterized as re-dispatch cost and equals 130 × 30 − 20 = $1300. When spread over total demand in the system, this equals $2.88/MWh above the uncongested price. The difference between the two approaches can become even more signiﬁcant when load is distributed across many locations. At ﬁrst glance, it would appear that under LMP the price paid by the load increases. However, this is not necessarily the case once FTRs are considered. In this example, the difference between total payments by the load and total payments to generators equals $7200. These congestion rents can be used to exactly fund FTR allocations. The total FTRs made available must be simultaneously feasible (e.g., 270 MW of FTRs from node 1 to 3 and 180 MW from node 2 to 3). If all FTRs are held by the load, the effective price paid by the load drops to $/24MWh which happens to be the weighted average of the cost of running the two generators at nodes 1 and 2. Ironically, the introduction of LMP with FTRs lowers the price paid by the load to even less than the uncongested case and equal to what might be expected in a pay-asbid auction. This example illustrates that using congestion rents as a metric may overstate the impact that congestion has on actual load payments. It is also interesting that although single price auctions and LMP markets are sometimes criticized for increasing price shocks for customers, the combination of LMP and FTR markets may actually move an organized market closer to a pay-as-bid market. Although FTRs were able to hedge congestion entirely in this example, this is not true in general. For example, if any out-of-merit generation were required at node 3, it would be paid the higher LMP and these payments could not be hedged by FTRs. Nevertheless, the example illustrates how the elimination of congestion may not always lower effective prices seen by consumers which could add yet another challenge to the already complex process of transmission expansion.

EBL

Transmission Markets, Congestion Management, and Investment

161

4.4. Congestion Metrics Under cost-of-service regulation, utilities made transmission and generation investment trade-offs as a part of their integrated resource planning (IRP). Under a market paradigm, generation and transmission investment decisions are often made by different parties that must rely on transparent access to market data and prices. One metric that plays a signiﬁcant role in transmission investment decisions is the magnitude of transmission congestion. Some of the metrics that have been used to measure congestion are as follows: • • •

Re-dispatch costs, Congestion rents or payments to FTR holders, Unhedgeable congestion costs.

Some metrics are directly observable from published LMPs while others require simulations of counter-factual scenarios as will be explained shortly. The term “re-dispatch” itself can be misleading when the system dispatch established in a day-ahead market already considers transmission constraints. There is inevitably some level of re-dispatch that must be performed in real time to account for constraints or changes in system conditions that were not already established. However, the term re-dispatch cost refers to the difference between the production cost implied by the actual system dispatch considering all transmission constraints and the production cost in a hypothetical scenario where no transmission constraints are binding. Although actual production costs of each generator may not be known, for the purposes of this metric the costs are assumed to equal offer prices submitted by generators. Since published LMPs are only calculated for the actual dispatch and they do not directly reveal offer prices, published market data does not lend itself to an easy calculation of re-dispatch costs. However, re-dispatch costs have traditionally been the metric of choice in transmission upgrade decisions. For example, the NYISO has used ofﬂine simulations to calculate re-dispatch costs. In 2003, re-dispatch costs in NYISO were estimated to be $85 million or approximately 1.5 per cent of total load billings. In 2004, the NYISO re-dispatch costs were estimated to be $72 million.23 These ﬁgures are smaller than what would be indicated by other metrics such as congestion rents. The congestion rent metric is widely used to report congestion and can be calculated from published market data. Ignoring the effect of losses, congestion rents are simply the difference between payments made by loads and exports and the revenues received by generators and imports. In the absence of congestion, all LMPs would be equal and congestion rents would be zero.24 They can also be calculated by adding up the products of the shadow prices and ﬂows across all congested lines. The shadow price on a line equals the change in dispatch costs for a unit change in the capacity of the line. Ignoring any impact of revenue shortfalls due to transmission derates, congestion rents would also equal the payments made to FTR holders. Figure 4.7 shows congestion rents in a few selected markets. In 2003, congestion rents in NYISO were approximately $688 million, signiﬁcantly higher than re-dispatch costs.

EBL

23

For more information on the New York ISO’s ongoing analysis of historic congestion costs, see http://www.nyiso.com/public/services/planning/congestion_cost.jsp 24 The incorporation of losses in LMPs would cause prices to vary even in the absence of congestion and result in a surplus similar to congestion rents. Unlike congestion rents that are used to fund FTRs, loss over-collections are refunded to grid users in some manner. Currently, NYISO and ISO-NE incorporate marginal losses in LMPs while PJM, MISO, and CAISO are working to implement them.

Competitive Electricity Markets

162

Total payments by load were approximately $6 billion. In PJM, congestion rents in 2005 were $2.09 billion. Total load billings were approximately $22.63 billion. Although total congestion costs in PJM increased signiﬁcantly since 2000, they have ranged between 6 and 10 per cent of the total load billings with 2005 representing the peak year. Congestion rents are directly observable from LMPs which can make them a transparent metric that is easy to cite. However, the caveat is that while LMPs and congestion rents can direct attention to where transmission investment is needed, additional information may be required to make actual investment decisions. Unhedgeable congestion refers to the impact of congestion on load that cannot be hedged by FTRs. If FTRs are allocated to load, a part of the increase in LMPs due to congestion can be hedged. However, not all of the price increase can be hedged. For example, load in an import region such as New York City could hedge the price increase due to congestion for all of its imports using FTRs (also known as Transmission Congestion Contracts or TCCs in NYISO). However, it could not use TCCs to hedge the price increase for purchases from local in-city generation. This creates what is characterized as unhedgeable congestion. The greater the presence of out-of-merit local generation in load pockets, the greater the level of unhedgeable congestion. In contrast, if there is no out-of-merit generation in a load pocket, there is no unhedgeable congestion. According to PJM, unhedgeable congestion in 2005 was roughly $103 million, which is signiﬁcantly smaller than the corresponding congestion rents. PJM’s deﬁnition of unhedgeable congestion also assumes any in-merit or economic local generation can serve as a hedge in addition to FTRs (Figs 4.7 and 4.8). At ﬁrst glance the congestion metrics described above increased from 2000 to 2005 in both PJM and NYISO but the same does not appear to be the case for California. However, a closer examination reveals that the CAISO also reported signiﬁcant intra-zonal congestion that was addressed using a variety of measures. In 2004, intra-zonal congestion was at its highest at approximately $426 million. It should also be noted that intra-zonal congestion costs are based on a re-dispatch cost metric and should not be compared to the congestion rent metric reported for inter-zonal congestion. Figure 4.9 shows intra-zonal congestion costs incurred under different categories of resources were used to resolve intra-zonal congestion in the CAISO markets. In contrast, inter-zonal congestion in the CAISO markets

EBL

Congestion costs

NYISO

PJM

CAISO

2000

$ (millions)

1500

1000

500

0 2000

2001

2002

2003

2004

2005

2006

Fig. 4.7. Congestion costs using the congestion rent metric across different markets. Increases in PJM congestion were partly attributable to the expansion of PJM’s footprint. Intra-zonal congestion for CAISO can be signiﬁcant and is shown separately. Sources: NYISO, PJM, and CAISO.

Transmission Markets, Congestion Management, and Investment

163

3.5

PJM congestion costs 3.0

$/MWh

2.5 2.0 1.5 1.0 0.5 0.0 2000

2001

2002

2003

2004

2005

2006

Fig. 4.8. PJM congestion rents on a per unit basis. Source: PJM State of the market reports.

CAISO Intra-zonal congestion 500

MILCC

RMR

RT Redispatch

EBL

400

$ (millions)

103 300

49

200

100

46 27

274

78 0

2003

2004

36

17

72

80

114

119

2005

2006

Fig. 4.9. CAISO intra-zonal congestion costs. Besides real-time re-dispatch, the CAISO also dispatches units under Reliability Must Run (RMR) contracts or under a must-offer requirement that requires paying minimum load cost compensation (MLCC). Source: CAISO.

decreased after expansion of Path 15, a major transmission interface between northern and southern California. Another metric that is sometimes cited as a measure of congestion is the frequency of curtailments made under TLR procedures. TLRs are administrative curtailment procedures used by transmission providers when scheduled transactions cannot be accommodated. According to NERC data, the number of TLRs has increased annually since 1997 (Fig. 4.10). The number of TLRs requiring curtailment of ﬁrm transmission ﬂows has also grown, from under 10 before 2001 to 70 in 2006. This increase is consistent with the rise in inter-regional

Competitive Electricity Markets

164 3000 2500 2000 1500 1000 500 0 1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Fig. 4.10. Level 2 and higher TLRs in the eastern Interconnection. Source: NERC TLR logs.

trade and higher levels of transmission congestion. It may be a little surprising that TLRs continued to increase despite the expansion of organized markets (e.g., PJM, MISO) in the eastern Interconnection. Although TLRs are typically not needed within organized markets due to the availability of sufﬁcient generation dispatch, they are used at the seams of organized markets, where power ﬂows from outside the organized markets may have a substantial impact on particular ﬂowgates. TLRs are not used in the western Interconnection, which uses a similar process known as the Unscheduled Flow Mitigation procedure. The Midwest ISO accounted for a signiﬁcant percentage of TLRs in 2004. In 2005 after the introduction of LMP, the number of TLRs decreased only slightly but the quantity of curtailment in MWh decreased signiﬁcantly. While increases in congestion metrics in LMP markets and the increase in TLRs both indicate increases in transmission congestion, the TLR trend is at least a little surprising since regions covered by LMP markets expanded during this period. Some of this increase could be a result of the increasingly complex seams created by expansion of some organized markets.

EBL

4.5. Financial Transmission Rights Presently, there are two major forms of transmission rights that exist in the United States – physical rights under Order 888 and ﬁnancial rights available in organized markets that incorporate LMP or its variants. As noted, the allocation of transmission capacity via daily auctions in organized markets enables greater certainty in access to the grid. However, this comes at the cost of increased price risk due to the differences in LMPs across locations during congestion. FTRs offer the ability to hedge price risk due to congestion by giving their holders a ﬁnancial revenue stream that can offset any congestion charges that apply (Hogan, 1992, 2002). In this sense, FTRs complement LMP markets to ensure that transmission service can be ﬁrm in both a physical and a ﬁnancial sense. However, not all LMP markets incorporate FTRs. For example, New Zealand, which was one of the ﬁrst international markets to implement LMP, has yet to implement any form of FTRs. Several design issues were addressed by stakeholders and FERC in the development of FTR markets. These include the very nature of the rights – e.g., should they be available

Transmission Markets, Congestion Management, and Investment

165

from any point or location in the grid to any other point or should they be restricted to physical interfaces or ﬂowgates? One concern, noted in passing earlier, was that the point-to-point nature of FTRs as initially conceived might prove too illiquid to support a major expansion in forward market trading. An initial attempt to solve this problem was the zonal pricing concept. A second type of proposal focused not on zones but on highly congested transmission facilities, which were called “ﬂowgates.”25 Proponents of ﬂowgate rights argued that if a small set of critical ﬂowgates could be deﬁned, it might allow for more liquid secondary markets that some considered essential for the efﬁcient allocation of FTRs (Chao and Peck, 1996; Chao et al., 2000). The debate was largely settled in favor of point-to-point rights although the simultaneous offering of ﬂowgate rights has been considered a possibility (O’Neill et al., 2002). Two markets that implemented zonal FTRs akin to ﬂowgate rights (because the zones were supposed to be on either “side” of highly congested paths) were California and Texas. Both markets are currently under redesign and will include point-to-point rights in the future.26 As noted, one concern expressed by market participants about point-to-point rights was that they could not be traded as easily in secondary markets. Such liquidity concerns were partially addressed by the development of trading hubs and zones. For example, FTRs can be purchased from a given generator in PJM to the PJM–West Hub. Similarly, FTRs can be purchased from one zone to another zone in NYISO. A signiﬁcant design feature of FTR markets is the manner in which FTRs are initially made available to market participants. The two main approaches are (a) to ﬁrst auction FTRs and then disburse the auction proceeds to some subset of market participants (e.g., to reduce access charges for all grid users), or (b) to ﬁrst directly allocate auction revenue rights (ARRs) or FTRs to Load Serving Entities (LSEs), and then hold an auction to allow for valuation of the rights and reassignment. The approach of holding the auction ﬁrst has been used at NYISO, ERCOT, and CAISO although that is changing under the market redesign.27 Figures 4.11 and 4.12 illustrate CAISO FTR auction results. In many instances revenues generated in FTR auctions far exceed the actual congestion or the revenues received by FTR holders. This has also been observed in FTR auctions in NYISO (Siddiqui et al., 2003). In some cases, ﬁrms purchasing FTRs also receive FTR auction revenues by virtue of being transmission owners. This needs to be considered in evaluating FTR auction results as a meaningful indicator of congestion. The hybrid approach that incorporates elements of auctions and direct allocations has become popular and its variants are in use at PJM, MISO, and ISO-NE. Under the hybrid approach, auction revenues are allocated to the LSEs that are holders of ARRs. Assuming that an LSE is allocated the appropriate set of ARRs, it can convert them into FTRs without worrying about the FTR auction price. The rules for allocating ARRs can vary across markets. In some markets such as MISO and PJM, the ARRs that an entity can request are linked to its historical loads and resources. In ISO-NE, an LSE may request an ARR that is linked not only to its load but also to every generator rather than a speciﬁc set of historical supply sources for that LSE. Such administrative rules for ARR allocation are

EBL

25

Note that this deﬁnition of ﬂowgates is not the same as the NERC deﬁnition used for purposes of monitoring facilities over which TLRs may be called. 26 The CAISO is expected to introduce LMP markets under its MRTU (Market Redesign Technology Update) program in 2008. 27 Under MRTU, the CAISO will directly allocate Congestion Revenue Rights (CRRs) to load serving entities based on their peak demand and historical supply sources.

Competitive Electricity Markets

166

FTR auction

CAISO FTR auctions

Congestion

200 Each year is April 1 - March 31 2006–07 congestion data is thru Sept

$ (millions)

150

100

50

0 2000–01

2001–02

2002–03

2003–04

2004–05

2005–06

2006–07

Fig. 4.11. CAISO FTR auction revenues and payments to FTR holders. With the exception of the initial auction, payments for purchasing FTRs have been less than congestion revenues. In some cases FTR auction prices may be unreliable as companies that bid for FTRs end up receiving FTR revenues as transmission owners. Source: CAISO.

2004 FTR auction revenue ($) 68,925 23,769,646

2004 FTR auction MW 355 Utilities

EBL

Merchants, Traders and retailers Munis

5,592

Utilities 5,544

Merchants, Traders and retailers Munis

77,499,873

Fig. 4.12. Distribution of CAISO FTR in 2004. Utilities in California still serve a majority of the load but hold roughly half the auctioned quantity of FTRs. However, utilities account for a much higher percentage of revenues paid in the FTR auctions. Source: CAISO.

necessary to preclude every market participant from requesting the same set of ARRs on a transmission path that is deemed valuable. However, even with administrative allocation rules, not all ARR requests between speciﬁc resources and loads may be feasible due to excessive nominations on congested transmission paths and some degree of pro-rationing on such paths may be necessary. Market participants that were accustomed to physical rights under Order 888 did not always adapt easily to the ﬁnancial rights approach in organized markets. One feature of physical rights that has been cited as useful was their “use it or lose it” property. In other words, holding a right that is not used does not cost anything although it may not be worth anything either. In contrast, FTRs provide revenues even if transmission service is not actually taken or does not actually match the FTR. However, FTRs can also become obligations if the direction of congestion changes. This has caused market participants to express a desire for option FTRs (Hogan, 2002). Figure 4.13 illustrates the properties of obligation and option FTRs. In general, issuing option FTRs reduces the total number of FTRs that can be awarded. Additionally, there are differences in the properties of the two types of FTRs. For example, an obligation FTR from node A to node B can be broken up

Transmission Markets, Congestion Management, and Investment

Payoff

167

Payoff {pb − pa}

max{0, pb − pa}

Pb-Pa

Obligations (pb − pa) = (phub − pa) + (pb − phub) (pb − pa) = −(pa − pb)

Pb-Pa

Options max(0, pb − pa) ≠ max(0, phub − pa) + max(0, pb − phub) max(0, pb − pa) = −max(0, pa − pb)

Fig. 4.13. Payoffs for obligations and option FTRs.

into an obligation FTR from node A to a hub and an obligation FTR from the hub to node B. This is not the case for option FTRs. Some markets such as PJM offer both option and obligation FTRs. Although the initial allocation of FTRs is done by ensuring that they are simultaneously feasible, there can be instances when subsequent derates in the transmission system cause some FTRs to be infeasible. In such instances, payouts for FTRs can be reduced or, alternatively, an uplift charge can be collected to ensure FTRs are fully funded once they have been allocated. Some markets such as PJM have yet to implement fully funded FTRs while others such as NYISO already do so. Assuming FTRs are fully funded, the allocation of uplifts to cover shortfalls and ensure full funding can become a contentious issue. One approach allocates the shortfall to transmission owners responsible for managing transmission outages. Critics of this approach argue that causes of the shortfall may extend beyond the responsibilities of the transmission owner. Figures 4.14 and 4.15 illustrate the magnitude of shortfalls in PJM and NYISO. With the exception of NYISO, FTR implementations have thus far been limited to annual rights. While congestion patterns and the contractual obligations of market participants change, the ability to readjust FTR positions annually is not necessarily a negative. However, uncertainty in being able to procure the desired quantities of FTRs has raised concerns about the lack of long-term FTRs. With a renewed focus on long-term contracting, attention has turned to making long-term FTRs available in all organized markets. The issue gained such prominence that there was a speciﬁc provision in EPACT 2005 that speciﬁcally required FERC to make long-term FTRs available in organized markets.28 FERC’s Order 681 speciﬁes a set of guidelines that must be met in implementing longterm FTRs (FERC, 2006b). In addition to requiring that the long-term FTRs have a sufﬁcient

EBL

28

Under EPACT 2005, a new Section 217(b)(4) of the Federal Power Act provides: “The Commission shall exercise the authority of the Commission under this Act in a manner that facilitates the planning and expansion of transmission facilities to meet the reasonable needs of load-serving entities to satisfy the service obligations of the load-serving entities to secure ﬁrm transmission rights (or equivalent tradable and ﬁnancial rights) on a long-term basis for long-term power supply arrangements made, or planned, to meet such needs.”

Competitive Electricity Markets

168 100.0%

80.0%

60.0%

40.0%

FTR Credit Percentage Planning Pd Percent 2003–04 97.7% 2004–05 100.0% 2005–06 90.7% 2006–07* 100.0%

Note: The monthly percentages include distribution of Excess Charges and may change for months in the 2006/07 planning period if excess charges are collected in the remainder of the planning period.

* through November 2006

20.0%

0.0% Jun- Sep- Dec- Mar- Jun- Sep- Dec- Mar- Jun- Sep- Dec- Mar- Jun- Sep03 03 03 04 04 04 04 05 05 05 05 06 06 06

Fig. 4.14. FTR payouts in PJM. Although there have been some months where the payouts for FTR holders have been reduced signiﬁcantly, the annual average payout has remained above 90 per cent. Source: PJM.

NYISO TCC Revenue shortfall 8000

TCC Payments

DA Congestion rents

EBL

7000

$ (millions)

6000 5000 4000 3500 2000 100 0 2002

2003

2004

2005

2006

Fig. 4.15. TCC payouts in NYISO. The difference between congestion rents and TCC payouts is collected using an uplift charge. Source: NYISO.

term to satisfy the needs of long-term contracting, the guidelines also touch upon many of the policy debates that occurred in designing annual FTRs. For example, long-term FTRs must be point-to-point, fully funded and available to LSEs without participating in an auction and be reassignable as load shifts between LSEs. The guidelines also require that long-term FTRs be made available for upgrades although this is one feature that is already available in most organized markets.

Transmission Markets, Congestion Management, and Investment

169

4.6. Transmission Investment Transmission can be both a complement to and a substitute for generation, i.e., transmission is necessary to serve load from distant generation but can also sometimes be substituted by building local generation. There are several major differences between transmission and generation. For example, a major driver behind the transition to competitive markets in generation was the argument that generation is no longer a natural monopoly with large economies of scale that requires construction of large power plants. In contrast, much of the transmission sector is still considered a natural monopoly with signiﬁcant economies of scale. There are some exceptions particularly with Flexible AC Transmission Systems (FACTS) and High Voltage DC (HVDC) technology that can lend themselves to merchant transmission projects. Another difference between transmission and generation is that transmission contributes a relatively smaller portion of retail electric bills compared to generation. This means that even signiﬁcant transmission investment may only amount to a relatively small increase in the ultimate cost to the retail customer and may lower the cost of generation resulting in a net saving.29 This has sometimes been used as an argument to build up the high-voltage transmission system much like the interstate highway system was. However, siting new transmission projects is not easy due to regulatory and environmental concerns. 4.6.1. Trends in transmission investment There are a few different dimensions to the issue of transmission investment. Perhaps the most signiﬁcant is the question of overall investment in transmission. Considerable attention has been given to declining investment in transmission infrastructure (Hirst, 2004; Joskow, 2005). As shown in Fig. 4.16, normalized transmission capacity in terms of MWmile per MW demand across all regions declined from 1989 through 2002. According to a survey by the Edison Electric Institute (EEI), overall transmission investment declined from 1975 to 1999 but has been increasing in recent years for both vertically integrated utilities and ITCs (EEI, 2005). Figure 4.17 illustrates in the increase in transmission investment based on an EEI survey of vertically integrated and stand-alone transmission companies. From 2000 through 2005, approximately $28 billion were invested in the US transmission system. From 2006 through 2009, an additional $31 billion is planned, representing a 60 per cent increase over the prior three-year period. More than 70 per cent of the roughly 150,000 miles of high-voltage transmission lines are estimated to be older than 25 years (Anderson et al., 2006). Declines in transmission investment have been attributed to several factors. These include a higher focus on generation with the creation of markets, challenges in transmission siting and obtaining rights of way, regulatory uncertainty, possible over-investment in earlier periods, and location of smaller generating units closer to load centers. At the same time, inter-regional usage of transmission and congestion has undoubtedly increased, putting greater pressure on the grid. One factor that may have contributed to declines in transmission investment was the focus on generation as that sector of the business was transitioning to a market paradigm. The higher pace of investment in generation relative to transmission was particularly

EBL

29

According to an FERC study on transmission congestion in 2001, transmission costs are roughly 6 per cent in contrast to generation costs of 74 per cent and distribution costs of 20 per cent in the average retail bill.

Competitive Electricity Markets

170

Normalized transmission capacity 250 200 150 100

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

0

1990

50

1989

MW-mi/MW demand

300

Fig. 4.16. Normalized transmission capacity across all NERC regions. Source: NERC, Hirst 2004.

7000

Transmission investment 6000

(2005$)millions

5000

EBL

4000 3000 2000 1000 0 2000

2001

2002

2003

2004

2005

Fig. 4.17. Actual transmission investment from 2000 to 2005. Source: EEI (2006).

evident during the generation boom period shown in Fig. 4.18. The potential of high unregulated returns on generation investments seemed more attractive than investing in regulated transmission, particularly when there were uncertainties on the structure of the transmission business. However, since the merchant generation meltdown and bankruptcies of many unregulated generation companies, investments in regulated transmission at incentive rates may not appear as unattractive for the levels of risk involved. Another factor in play was the regulatory uncertainty with respect to transmission which has since improved with greater clarity on transmission structures. From a fundamental’s perspective, even though inter-regional trade has increased signiﬁcantly with the creation of markets, new smaller generating plants can now be located closer to load centers than older larger base-load units. The availability of adequate transmission can sometimes make generation markets more competitive and has been cited as a factor that

Transmission Markets, Congestion Management, and Investment

171

Percent change over year end 1998 20%

Percent

15%

Generation plant construction (MW)

10%

Transmission investment ($)

5%

0% 1999

2000

2001

2002

Fig. 4.18. Difference between generation and transmission investment.

should be considered in expansion decisions (California ISO, 2004). All these factors need to be considered together in determining the optimal level of transmission investment going forward. As explained earlier in the chapter, simply quoting congestion rent metrics in LMP markets may not always be the right predictor of new transmission needs. In some cases, transmission cost allocation methodologies can also create opposition to transmission expansion proposals. For example, the expansion of a transmission bottleneck may increase prices in an exporting region. If costs of the expansion are allocated on a regional basis, customers in the exporting region may feel they have to pay for price increases. Investment trends in transmission may be further differentiated between ITCs and vertically integrated utilities. Proponents of ITCs present several arguments as to why ITCs may be better positioned than vertically integrated utilities when it comes to transmission investment. For example, an ITC has a singular focus on transmission, which eliminates competition between generation and transmission. They argue that an ITC does not have the perverse incentive to maintain congestion for protecting generation market share. According to the International Transmission Company, the Midwest ISO’s most recent transmission expansion plan has ITCs accounting for approximately 54 per cent of the $2.9 billion of transmission investment projected through 2009. The average ITC investment is at about seven times that for a non-ITC in MISO. Proponents of ITCs offer this statistic as evidence of their arguments on why ITCs may be better positioned than vertically integrated utilities to focus on transmission investment. Transmission planning has also witnessed signiﬁcant changes in the transition to competitive markets. Historically, transmission and generation investment were often considered substitutes and lumped together under integrated resource planning by utilities. The separation of transmission and generation ownership has made this task more challenging. At the same time, there have also been positive changes as the larger footprints of RTOs and ITCs make the task of regional transmission planning easier. Some degree of coordination between generation and transmission investment is also captured in the generator interconnection procedures used by transmission organizations. For example, any new generator in PJM needs to satisfy certain deliverability criteria and

EBL

172

Competitive Electricity Markets

is responsible for transmission upgrades that may be needed to satisfy this requirement.30 Such efforts are only now beginning to reach a level of maturity that can make it possible to accurately assess how well they are functioning (California ISO, 2006; PJM, 2006; ISO-NE, 2004 ISO-NE, 2006; NYISO, 2006).

4.6.2. The merchant transmission model Although most transmission investments to date have been made under a regulated cost of service approach, there are also so-called merchant or market-based transmission projects (Rotger and Felder, 2001). Merchant transmission projects rely on market rates for transmission service negotiated directly with purchasers of their capacity, and assume (along with the purchasers of their capacity) all of the market risk for their facilities. Transmission projects can often be lumpy and exhibit large economies of scale that do not always lend themselves to a merchant investment model, though there are exceptions such as FACTS and HVDC technology. Some AC projects using variable frequency transformers may also qualify. While it is typical for merchant transmission projects to involve an open season for their capacity, in theory a project could also be truly merchant and simply offer its capacity to users at market-based rates. In organized RTO markets that use ﬁnancial rights, FTRs or ARRs are created by a merchant project and some analysts thought initially that such rights would provide a signiﬁcant component of the project’s forecast revenue stream (with the remainder coming from bilateral contracts to reserve capacity on the transmission facility). However, in case of a lumpy investment, the award of FTRs may have little value if congestion is entirely eliminated. In principle, transmission merchant transmission capacity could be treated as dispatchable in the same manner as generation is dispatchable, i.e., portions of a merchant transmission line would become available only if the price spread across the line exceeds an offer price (O’Neill et al., 2005). The dispatchable transmission paradigm can also be used to relax normal line ratings for short periods under emergency conditions (e.g., it may be less expensive to temporarily expand a line rating and compensate its owner for a short interval instead of re-dispatching more expensive generation). However, reasonable rules for market power mitigation could prove a challenge before any such proposals are implemented. In addition to price signals and incentives for transmission investment, there are other factors that inﬂuence transmission investment decisions. These factors include transmission pricing and cost allocation, inter-regional planning, and challenges in siting and obtaining rights of way. Despite the presence of these various factors that make a direct correlation between LMP implementation and transmission investment difﬁcult, there are several instances where market prices have played a role in encouraging transmission projects. Speciﬁc examples in the northeast are the Neptune and Cross-sound cable projects that have helped expand the transmission capability into Long Island in

EBL

30

PJM determines the sufﬁciency of network transfer capability through a series of “deliverability tests.” This includes the Capacity Emergency Transfer Objective (CETO) and Capacity Emergency Transfer Limit (CETL) tests to check if energy from the aggregate of PJM resources can be delivered to capacity deﬁcient sub-regions within PJM. CETO represents the amount of energy a sub-area must be able to import in order to comply with reliability standards. The test is passed if the actual CETL exceeds the CETO. Generators must satisfy the deliverability test in order to be classiﬁed as “capacity resources.” PJM’s Reliability Pricing Model (RPM) has taken further steps to make generation capacity planning assumptions consistent with those used in transmission planning.

Transmission Markets, Congestion Management, and Investment

173

Table 4.7. Merchant transmission projects Project

Description

Expected completion

Neptune

600 MW HVDC submarine cable linking New Jersey (PJM) and Long Island (NYISO)

Completed in summer 2007

Cross Sound Cable

330 MW HVDC submarine cable linking Connecticut (ISO-NE) and Long Island (NYISO)

Completed in 2002, ﬁrst operated in 2003

Montana-Alberta Tie

300 MW, 230 kV line linking Lethbridge, Alberta and Great Falls, Montana

Open season in 2005

Seabreeze Power

550 MW HVDC submarine cable linking Vancouver Island and Port Angeles, WA.

Open season in 2005

Northern Lights

Three 3000 MW HVDC links connecting Alberta to Paciﬁc northwest and Wyoming/Montana to the southwest

Proposed in 2005, expected to be fully committed in 2009

Linden Variable Frequency Three 100 MW VFTs connecting generation at Successful open season Transformer Linden in PJM to the Goethals substation in results in 2007, Staten Island in New York City operation expected in December 2009

New York.31 Although the CAISO has yet to implement full LMP, it has had zonal prices since 1998 that reﬂected signiﬁcant congestion and inﬂuenced the expansion of a major transmission bottleneck called Path 15.32 Another example of merchant transmission investments includes the Sea Breeze proposal to build a 550 MW undersea cable between Vancouver Island and Port Angeles, Washington State. Table 4.7 lists major merchant projects that have been proposed or implemented so far.

EBL

31

The Neptune Project involves the installation of an approximately 600 MW (500 kV) fully controllable high-voltage, 65-mile direct current (HVDC) submarine electric transmission cable connecting power generation resources in New Jersey to electricity consumers in Long Island. Construction began in 2005 and the project was completed in summer 2007. The Cross Sound Cable (CSC) is a 24-mile, 330 MW high-voltage direct current (HVDC) submarine transmission system linking the electricity grids of New England and Long Island. The project was completed in 2002 but operations started only after the August 14, 2003 Blackout on an emergency basis, stopped again after protests by the Connecticut Attorney General and ﬁnally started after an agreement was reached in 2004. The CSC is currently owned by Babcock and Brown Infrastructure, an infrastructure investment fund based in Australia. The entire capacity for both projects is under contract to the Long Island Power Authority (LIPA). 32 Path 15 is a system of three 500 kV transmission lines that narrows down to two lines for an 83-mile stretch through California’s Central Valley. The Path 15 upgrade project added a third line to that segment and upgrades several major transmission substations that are part of Path 15. Completed in December 2004, the upgrades added 1500 MW of transmission capacity to the existing 3950 MW capacity between southern and northern California. The project involved a consortium between independent transmission entity Trans-elect, Paciﬁc Gas and Electric (PG&E) and the Western Area Power Administration. The California energy crisis highlighted the need to relieve congestion caused by Path 15 (estimated by the CAISO at $221.7 million in additional energy costs for consumers between September 1, 1999 and December 31, 2000.) In May 2001, Secretary of Energy Spencer Abraham directed Western to explore an upgrade and to determine whether non-Federal parties would be interested in helping ﬁnance and co-own the system additions. The partnership with Trans-elect and PG&E resulted from this effort.

174

Competitive Electricity Markets

4.6.3. Factors inﬂuencing regulated transmission investment and other developments Cost allocation for non-merchant transmission projects can also be important in investment decisions. The approaches range from postage stamp to license plate, rolled-in socialization to more targeted allocations based on beneﬁts of projects and hybrid approaches that socialize costs of high-voltage facilities but use more targeted approaches for local investments.33 Cost allocation for transmission needed for renewable resources such as wind are also evolving.34 The challenges in transmission investment were recognized by Congress in EPACT 2005 and there are several new initiatives underway to address this issue (the major elements are listed in Box 4.2) These include backstop siting authority for FERC, the designation of National Interest Electric Transmission (NIET) Corridors, and new incentives for transmission under a recently issued FERC rule. The US Department of Energy completed an initial study in October 2006 that found two areas of critical congestion – southern California and the Atlantic coast from New York City to northern Virginia. These areas were the basis for formal designation of two NIET corridors in April 2007. The two areas are the Mid-Atlantic area corridor and the southwest area corridor.35 EPACT 2005 created for the ﬁrst time federal eminent domain for electric transmission although this authority is supplemental to and does not preempt state law except in four narrow situations. FERC can permit and condemn right-of-way for construction of an electric transmission line provided that: •

the state where the line is to be located lacks siting authority or state law prohibits siting to achieve interstate beneﬁts; • the permit applicant is not eligible for site approval because it does not provide retail service in the state; • the state with siting authority fails to act on an application within one year; or • the state siting body attaches conditions that will prevent congestion reduction or make the new line economically infeasible.

EBL

Procedures for permitting construction of transmission lines in NIET corridors are described in Order 689 (FERC, 2006d). Section 1241 of EPACT 2005 added a new section 219 to the Federal Power Act that required FERC to establish incentive-based rate treatments that promote capital investment in transmission of electric energy in interstate commerce, regardless of the ownership of the facilities; provide a return on equity (ROE) that attracts new investment in transmission facilities (including related transmission technologies); encourage deployment of transmission technologies and other measures to increase the capacity and efﬁciency of existing transmission facilities and improve the operation of the facilities. Order 679 modiﬁed by Order 679A fulﬁlls that directive (FERC, 2006a). Order 679 establishes a higher incentive-based ROE for investments that enhance reliability and/or reduce congestion. The rule also offers incentive-based rate treatments to encourage ITC/Transco formation and membership in RTOs/ISOs or other approved transmission organizations. 33

See, e.g., 119 FERC ¶ 61,063 Opinion No. 494, PJM Interconnection, L.L.C. Docket Nos. EL05-121-000 and EL05-121-002, 19 April 2007. 34 See 119 FERC ¶ 61,061 California ISO, Docket EL07-33-000, Order granting petition for declaratory order, April 19, 2007. 35 Further details on the NIET corridors can be found at http://nietc.anl.gov/indexnew.cfm

Transmission Markets, Congestion Management, and Investment

175

Box 4.2 Elements of Energy Policy Act of 2005 Major elements of the Energy Policy Act of 2005 related to Transmission • • • • • •

Creation of Incentives for Transmission Investment – FERC Order 679 Creation of long-term transmission rights in organized markets – FERC Order 681 Repeal of Public Utility Holding Company Act (PUHCA) Designation of National Interest Electric Transmission Corridors Federal backstop siting authority for electric transmission – FERC Order 689 Mandatory Standards for Reliability

The rule also contains other incentives such as allowing 100 per cent Construction Work in Progress (CWIP) in rate-base, recognizing the long lead times involved in transmission projects. The rule does not dismiss the recovery of acquisition premiums in incentive rates but suggests a case-by-case approach to evaluate such proposals. Order 679A reafﬁrmed the proposals in Order 679 but added certain clariﬁcations. For example, the review process for a project must speciﬁcally show that the project is needed for reliability or reduces congestion. FERC would also not routinely grant incentive rates at the top end of the zone of reasonableness and applicants would need to justify a higher ROE. The major elements of Order 679 are listed in Box 4.3.36

EBL

Box 4.3 Key Provisions of FERC Order 679 Key provisions of FERC Transmission Incentives Rule Order 679 • • • • • • • •

Incentive rates of return on equity for new transmission investment Full recovery of prudently incurred construction work in progress Full recovery of prudently incurred pre-operations costs Full recovery of prudently incurred costs of abandoned transmission facilities Use of hypothetical capital structures Accumulated deferred income taxes for stand-alone transmission companies Accelerated depreciation Deferred cost recovery for utilities that cannot pass-through costs of new transmission investments due to retail rate freezes • Higher rates of return on equity for utilities that join and/or continue to be members of transmission organizations, including (but not limited to) RTOs and ISOs

36

Several transmission projects have been granted incentive rates under Order 679, see e.g., Baltimore Gas and Electric Docket No ER07-576-000, 121 FERC 61, 167 and Southern California Edison, Docket No EL07-62-000, 121 FERC 61, 168 November 17, 2007.

176

Competitive Electricity Markets

Finally, the Public Utility Holding Company Act (PUHCA) of 1935 was enacted by Congress in response to the collapse of a few large companies that controlled signiﬁcant generation and transmission assets during the Great Depression. PUHCA split up the companies and imposed restrictions on utilities operating in more than one state. It limited US electric utilities to a single geographic region in which integrated operation was feasible and prohibited utility ownership of businesses that were not reasonably incidental or functionally related to the utility business. PUHCA was viewed by many as impeding investment in transmission companies in more than one region. Under PUHCA, mergers and acquisitions by holding companies were conﬁned to interconnected areas or regions although “virtual interconnection” arguments have been used at times to justify mergers of companies not physically contiguous. The geographic requirement of PUHCA was considered counter-intuitive with respect to addressing generation market power by encouraging mergers between adjacent markets. PUHCA also restricted non-utilities from directly owning and controlling utilities. With its repeal, private equity and other ﬁnancial investors are expected to enter the sector more easily. This is a signiﬁcant development given the increasingly active role played by private equity.

4.7. Conclusions After a decade of transmission open access, transmission markets in the United States appear to have reached a certain level of maturity. Two distinct models co-exist in different parts of the country: Order 888-based markets and organized RTO markets. Although organized RTO markets were initially seen as an evolution of the older Order 888 regime, there is recognition that the two are likely to continue to co-exist for some time. Consequently, there have been efforts to further reform Order 888-based markets within that paradigm. There are also some RTO markets that have chosen to operate Order 888 markets. There is recognition that despite the many advantages that organized markets might offer, they too have room for improvement such as the development of long-term transmission rights. On the structural and investment front, the early debate on whether for-proﬁt or nonproﬁt grid operators made more sense has been largely replaced by the question of independent transmission companies versus vertically integrated utilities. Overall investment in transmission, which suffered in comparison to generation for many years, appears to have turned a corner. The growth of RTO footprints and maturity in planning processes and incentives in EPACT are all likely to be positive inﬂuences in this area. Investments under the merchant transmission model have been somewhat limited. There have been signiﬁcant increases in transparent market data on congestion albeit there are many different metrics in use across regions. Even though there is some degree of consensus that transmission investment has not kept pace with the growth in generation and demand, using the wrong congestion metric can artiﬁcially inﬂate the levels of transmission investment that may be required.

EBL

4.8. Acknowledgment The author would like to thank Udi Helman and Perry Sioshansi for providing useful comments.

Transmission Markets, Congestion Management, and Investment

177

References Anderson, K., Furey, D., and Omar, K. (2006). Frayed wires: US transmission system shows its age. Fitch Rat., Special Report, 25 October. Adib, P. and Zarnikau, J. (2006). Texas: The most robust competitive market in North America. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Elsevier. California ISO (2004). Transmission Economic Assessment Methodology (TEAM), June. California ISO (2004a). Annual Report on Market Performance, 2005. California ISO (2006). 2005 Annual Report on Market Performance, April 2006. Casazza, J. (1993). The Development of Electric Power Transmission. IEEE Press. Chandley, J.D. and Hogan, W.W. (2002). Independent transmission companies in a regional transmission organization. Harvard University, Cambridge, MA, 8 January. Chao, H. and Peck, S. (1996). A Market mechanism for electric power transmission. J. of Reg. Econ., 10(1), 25–60. Chao, H., Peck, S., Oren, S., and Wilson, R. (2000). Flow-based transmission rights and congestion management. Elec. J., October, 38–58. Department of Energy (DOE) (2006). National Electric Transmission Congestion Study Report, August. Edison Electric Institute (EEI) (2005). EEI Survey of Transmission Investment, May. Federal Energy Regulatory Commission (FERC) (1996). Promoting Wholesale Competition through Open Access Non-discriminatory Transmission Services by Public Utilities; Recovery of Stranded Costs by Public Utilities and Transmitting Utilities. Order No. 888, 61 FR 21540, 10 May. Federal Energy Regulatory Commission (FERC) (1996a). Open Access Same-Time Information System. Order No. 889, 61 FR 21, 737, 10 May. Federal Energy Regulatory Commission (FERC) (1999). Regional Transmission Organizations. Order No. 2000, Docket RM-99-2-000, FERC ¶ 61,285, 20 December. Federal Energy Regulatory Commission (FERC) (2006). Promoting Investment Through Pricing Reform. Order 679, Docket No. RM06-4-001, FERC ¶ 31,222, 31 July. Federal Energy Regulatory Commission (FERC) (2006a). Long-Term Firm Transmission Rights in Organized Electricity Markets. Order No. 681, Docket No. RM06-8-001, FERC ¶ 31,226, 1 August. Federal Energy Regulatory Commission (FERC) (2006b). Rules Concerning Certiﬁcation of the Electric Reliability Organization; Procedures for the Establishment, Approval and Enforcement of Electric Reliability Standards. Order No. 672, 71 FR 8662, 17 February. Federal Energy Regulatory Commission (FERC) (2006c). Mandatory Reliability Standards for the Bulk Power System. (Docket No. RM06-16-000), 20 October. Federal Energy Regulatory Commission (FERC) (2006d). Regulations for Filing Applications for Permits to Site Interstate Electric Transmission Facilities. (Docket No. RM06-12-000), 16 November. Federal Energy Regulatory Commission (FERC) (2007). Preventing Undue Discrimination and Preference in Transmission Service. Order No. 890, Docket Nos. RM05-25-000 and RM05-17-000, 16 February. Federal Energy Regulatory Commission (FERC) (2007a). Mandatory Reliability Standards for the Bulk Power System. Order No. 693, Docket No. RM06-16-000, 17 March. Federal Energy Regulatory Commission (FERC) (2007b). Order on Violation Risk Factors. Docket No. RR07-9-000 and RR07-10-000, 18 May. Harvey, S. and Hogan, W. (2002). Loss hedging ﬁnancial transmission rights. In School of Government (J.F. Kennedy ed.) Harvard University, Cambridge, MA, 15 January. Henney, A. and Russell, T. (2002). Lessons From The Institutional Framework Of Transmission, System Operation, and Energy Markets In Most West European Countries and Some Other Countries – The Case for Transcos. Docket No. RM 01-12-00, Federal Energy Regulatory Commission. Hirst, E. (2004). U.S. Transmission Capacity: Present Status and Future Prospects. Edison Electric Institute, June. Hogan, W.W. (1992). Contract networks for electric power transmission. J. of Reg. Econ., 4(3), 211–42. Hogan, W.W. (2002). Financial Transmission Right Formulations Report. Harvard University, Cambridge, MA, March 2002. ISO-New England (2004). Regional Transmission Expansion Plan (RTEP04) Technical report, 21 October.

EBL

178

Competitive Electricity Markets

ISO-New England (2006). 2006 Regional System Plan. October. Joskow, P. (2005). Patterns of transmission investment. MIT, March. NERC (2006). Long Term Reliability Assessment. http://www.nerc.com/∼ﬁlez/rasreports.html NYISO (2006). Comprehensive Reliability Plan. August. O’Neill, R., Helman, U., Hobbs, B., Benjamin, F., Stewart, Jr., W.R., and Rothkopf, M.H. (2002). A joint energy and transmission rights auction: Proposal and properties. IEEE Trans. on Pow. Sys., 17(4), 1058–67. O’Neill, R., Baldick, R., Helman, U., Rothkopf, M., and Stewart, Jr., W.R., (2005). Dispatchable transmission in RTO markets. IEEE Trans. on Pow. Sys., 20(1), 171–9. PJM (2006). Regional Transmission Expansion Plan. February. PJM (2006a). Annual State of the Markets Report, April. Rajaraman, R. and Alvarado, F. (1998). Inefﬁciencies of NERC’s transmission loading relief procedures. Elec. J., October. Rotger, J. and Felder, F., (2001). Promoting efﬁcient transmission investment: The role of the market in expanding transmission infrastructure. November 2001. Rudkevich, A, Hausman, E., Tabors, R., Bagnall, J., and Kopel, C. (2005). Loss hedging rights: A ﬁnal piece in the LMP Puzzle. Proceedings of the 38th Hawaii International Conference on System Sciences. Ruff, L. (2002). Deﬁning and allocating RTO functions. FERC Technical Conference on RTO Functions and Characteristics, Washington DC, 19 February. Ruff, L. (2006). The missing markets: What they are (not) and what (if anything) to do about it. Harvard Electricity Policy Group, La Jolla, California, 2 March. Siddiqui, A.S., Bartholomew, E., Marnay, C., and Oren, S.S. (2003). On the efﬁciency of the New York Independent System Operator market for transmission congestion contracts. Working Paper, University of California, Berkeley. Singh, H. (2005). Incentives for transmission. RTO Governance Conference, Washington DC, 31 March. Sweeney, J. (2006). California electricity restructuring, the crisis, and its aftermath. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Elsevier. Wood, A.J. and Wollenberg, B. (1996). Power Generation, Operation and Control. New York: John Wiley and Sons.

EBL

Chapter 5 The Design of US Wholesale Energy and Ancillary Service Auction Markets: Theory and Practice UDI HELMAN,1 BENJAMIN F. HOBBS,2 AND RICHARD P. O’NEILL1 1

Federal Energy Regulatory Commission, Washington, DC, USA; Department of Geography and Environmental Engineering, The Johns Hopkins University, Baltimore, MD, USA

2

Summary

EBL

In the United States, after about a decade of experience with market design, wholesale spot markets operated by Independent System Operators (ISOs) around the country have largely converged on core design elements. This chapter provides a detailed description of how these markets operate. In particular, most of these markets have day-ahead and real-time auction markets for energy and certain ancillary services, typically regulation and operating reserves. The energy auctions can accommodate both physical and virtual supply offers and demand bids. In the regulation and reserve auctions, only physical offers are currently allowed, including those from dispatchable demand. With the submitted day-ahead offers, bids, and non-price schedules, the ISO conducts a security-constrained unit commitment auction, which selects the generation units that will run for every hour of the day subject to all relevant unit and transmission network constraints. For the energy markets, the auction outcome is two sets of prices that together clear the market: locational marginal prices (LMPs) for energy, which include congestion and loss components, and separate payments to ensure revenue sufﬁciency for any offer or bid costs, such as generation start-up costs, not recovered through LMPs. The real-time energy markets have also progressively incorporated most elements of this design, although auction procedures are somewhat different from the day-ahead market. Integrated into the day-ahead and real-time energy markets are markets for regulation and operating reserves, co-optimized with energy. Market prices for these ancillary services typically incorporate an opportunity cost payment with respect to any foregone energy sales as well as availability payments, if needed. As with energy, revenue sufﬁciency is guaranteed through additional payments. To provide insight into each stage of the market and into the principles of locational marginal pricing, the chapter provides 179

180

Competitive Electricity Markets

a simple numerical example of an energy auction on an electricity network. Finally, the chapter brieﬂy explores other key design issues, such as refunds of surplus marginal congestion and loss payments, market power monitoring and mitigation, addressing continuing market seams, software development, and extensions of the market design.

5.1. Introduction In the UNITED STATES, wholesale markets for electric power have evolved along two basic organizational approaches, both consistent with the open access transmission regime established by the US Federal Energy Regulatory Commission (FERC) in 1996 (FERC, 1996a). In the ﬁrst type of market, electric utilities and non-utility generators contract bilaterally among themselves for energy on a forward basis. The utilities that own the transmission facilities determine the available quantity and price of transmission access and physical scheduling rights, subject to open access rules. While private power exchanges may form to facilitate forward energy contracting, there is no co-ordinated spot energy market that encompasses the territory of multiple utilities and heretofore no price-based congestion management. In the second type of market, an independent third-party entity, which for purposes of this chapter will be called an Independent System Operator (ISO), operates organized regional bid-based auction markets for spot energy, various types of ancillary services, and possibly capacity, and allocates all transmission capacity and transmission property rights in an efﬁcient and non-discriminatory fashion. Spot transmission usage is subject to charges for congestion and losses. These ISO markets are the subject of this chapter. The chapter will focus on the design of the daily energy, regulation, and operating reserves markets. The chapter will not review ISO market performance, except on occasion to explain a design decision.1

EBL

5.1.1. Overview of Market Design The ISO market designs that have arisen in different regions of the United States have similarities and differences, resulting from the fairly high degree of regional decisionmaking in the regulatory reform of the US electricity sector.2 In general, the ISO operates a day-ahead market and a real-time, or dispatch, market. These are organized as sealedbid, multiple-unit auction markets. The day-ahead market is a forward market in which accepted offers or bids can choose not to perform in real-time (i.e., go to physical delivery) as long as they buy back or sell back their positions. Put another way, the real-time market determines the prices of “deviations” from the day-ahead schedule. Due both to ﬁnancial incentives and certain administrative rules discussed below, most accepted day-ahead offers and bids that reﬂect physical supply and demand do go to physical delivery and hence the two markets collectively can be thought of as the “spot” market. In both markets, suppliers submit offers prior to a trading deadline (usually the prior morning for the day-ahead market and about 1 hour before the real-time market). These 1 Market performance is discussed extensively in the US ISO annual state of the markets reports, e.g., PJM 2006a, Potomac Economics 2003–2006 (for New York ISO). In addition, there is a large research literature on this subject. 2 The evolution of US federal rulemaking, including the failed effort to standardize the market designs, is discussed in O’Neill et al. (2006) and O’Neill and Helman (2007).

Design of US Wholesale Energy and Ancillary Service Auction Markets

181

offers must usually specify the minimum and maximum MW that can be produced by the generator, the price of energy ($/MWh) over the range of its available output, a start-up cost ($), a no-load cost ($), and a number of physical characteristics, such as how rapidly the generator can increase or decrease output (called the “ramp rate” and measured in MW/minute). For operating reserves and regulation, suppliers typically provide additional price offers and physical parameters to deﬁne capability. Both the day-ahead and real-time auctions conduct “security constrained unit commitment,” in that they specify exactly which generation units should be turned on in each hour, their level of output, and the length of time they should run over the day, based on start-up and energy offer prices and the other ﬁnancial and physical parameters and transmission network constraints. In the day-ahead market, the unit commitment decision is integrated into the auction. In real-time, unit commitment is conducted through a parallel pre-auction program that “looks-ahead” based on a forecast to determine which units to commit or decommit, while the auction function only adjusts the output of already committed units on a 5–15 minute basis. In addition, because energy, regulation, and reserves can be either or both complementary and substitute uses of a generator, establishing the correct auction constraints to reﬂect these possible relationships, both day-ahead and real-time, has proven to be important for purposes of economic efﬁciency. For both day-ahead and real-time markets, the auction result is a market clearing with uniform energy prices at each speciﬁed commercial location (nodes or buses) on the transmission network, called “locational marginal prices” (LMPs). Buyers pay the LMPs at their locations (which to date have typically been zonal prices comprised of a load-weighted average of the LMPs in the zone) and sellers are paid the LMPs at their locations. A key function of the energy auctions with LMP is to put a market price on marginal transmission usage, including congestion and losses. In the day-ahead market, congestion is managed instantaneously as part of the auction optimization, which respects most relevant transmission constraints. In the real-time market, the system operators manage congestion on a minute-to-minute basis in part through auction prices and in part through non-market operating decisions. Regulation and operating reserve prices are calculated slightly differently, typically incorporating an opportunity cost payment and an availability price offer into the calculation of market clearing prices, which are determined for pre-deﬁned zones. Additional “pay-as-bid” payments are made to accepted offers and bids in any of these markets if their auction revenues do not fulﬁll their offer or bid terms (e.g., if a generator is started up but then does not run for sufﬁcient hours to cover the sum of its offer prices for start-up, no-load, energy, regulation, or reserves). These vast regional wholesale spot markets, several consisting of tens of thousands of simultaneously determined prices at locations on the grid, are one of the signal technological achievements to date of the regulatory reform of the US electricity industry. The ISO market designs are also serving as models for other countries’ attempts at market development. At one level, they operate very well, allowing for the regional market operator to capture efﬁciencies made possible through large-scale optimization. Buyers can schedule their own resources (owned or contracted) or buy spot. Sellers can optimize whether to fulﬁll their forward contracts with their own resources or through the spot market and can offer any residual capacity into the spot market. The record of the ﬁrst decade of energy and ancillary service market design in the US ISOs suggests that while some conceptual design issues have largely been resolved, and there have been many technological innovations and advances in auction design, there remain many design and implementation challenges. First, these markets remain “incomplete” in the sense that they do not price adequately all generator services and

EBL

182

Competitive Electricity Markets

physical requirements associated with power transactions (transmission markets present other types of incompleteness). Moreover, while sellers have sufﬁcient ﬂexibility to offer their true costs, buyers are not sufﬁciently price responsive, due primarily to regulatory and technological barriers. Second, the ISO markets, while fairly unconcentrated in the aggregate, are not perfectly competitive at all locations and at all times due to transmission constraints. The primary method to control intermittent market power in the ISO markets has been through supply offer price caps. But offer caps suppress market prices and sometimes hinder investment. As a result of these two problems, additional pricing rules, such as the “scarcity pricing” discussed in this chapter, and additional markets established through regulatory requirement, such as capacity (or resource adequacy) markets, have been perceived as necessary to support both power system reliability and economic efﬁciency. Some of these design features can be removed (or become irrelevant) as technology evolves and the markets become more complete. With respect to several design elements, no deﬁnitive best practice has yet emerged. To illustrate this, in several sections of the chapter, the rules and procedures in two US ISO markets, PJM and New York, are discussed. These markets were picked because they offer useful comparisons on several design choices in the auction market rules, the integration of market and system operations, and market power mitigation. Nevertheless, the differences, while worthy of consideration, should not obscure the convergence of the US ISO markets over time on important common design elements. 5.1.2. Chapter organization

EBL

In an earlier survey (O’Neill et al., 2006), electricity market design principles were discussed at a relatively high level and US ISO market features were reviewed. The complementary purpose of this chapter is to provide an in-depth description of the sequence of the daily ISO auction markets for energy and certain ancillary services. There is sufﬁcient detail to allow the reader to track most of the market rules and computational procedures that affect the ultimate price of wholesale power transacted through such markets. The chapter is organized as follows. Section 5.2 provides an overview of electricity market design choices and options for market power mitigation. Sections 5.3–5.5 discuss the auction sequence of the day-ahead market, the reliability unit commitment, and the real-time market. A numerical example is introduced in Section 5.3 to motivate the auction description, and continues in most subsequent sections. A mathematical statement of the auction model used in the example is provided in Appendix 5A. Section 5.6 examines revenue sufﬁciency guarantees that support efﬁcient market behavior in each step of the auction sequence. Section 5.7 explains how surplus spot energy payments collected by the ISO due to pricing of congestion and losses are refunded to market participants. Section 5.8 discusses market power monitoring and mitigation in these auctions. Section 5.9 collects several additional topics in ISO market design and implementation, including the ISO’s longer-term markets and operational functions. Section 5.10 discusses possible next stages in the design of these markets. Section 5.11 presents conclusions. 5.2. The Development of Wholesale Energy Auction Market Designs The designs of the wholesale auctions for energy and ancillary services described in this chapter emerged over a decade of research and experience in the United States and elsewhere and continue to be modiﬁed and reﬁned [see, e.g., the surveys in O’Neill et al. (2006); FERC (2002); Stoft (2002); Wilson (2002)]. This section brieﬂy reviews

Design of US Wholesale Energy and Ancillary Service Auction Markets

183

some of the alternative designs that were considered in that period and explains why certain choices appeared desirable for theoretical, practical, and regulatory reasons. These include • • •

deﬁnition of the market products, functions of the ISO, choices in electricity auction design, including the question of uniform versus discriminatory pricing, • sequencing of markets and reliability functions, • market power monitoring and mitigation, and • scarcity pricing.

5.2.1. Deﬁnition of market products A ﬁrst step in designing markets is to deﬁne the products, including the time and location at which the transaction takes place. While in most commodity markets, product deﬁnition is a matter for private ﬁrms to determine, in the electricity spot markets operated by ISOs, the nature of the technology in real-time (e.g., balancing requirements and lack of storage) and the presence of reliability requirements has led to a strong regulatory role in product deﬁnition (e.g., in FERC, 1996a), although ISOs have had latitude to adjust some deﬁnitions and parameters to ﬁt their systems. Energy is the primary wholesale product traded in the ISO electricity markets (as measured in terms of quantities and monetary value), and it is deﬁned straightforwardly as mega-watt-hours (MWh) injected or withdrawn at a location or locations (e.g., hub or zone) on the transmission network in one of the hourly markets (day-ahead or realtime). For purposes of spot energy market design, sales of energy are typically limited to generator output between a unit’s minimum and maximum operating levels. Price offers for energy are typically required to be linear (see Table 5.7). Generator “start-up,” i.e., the short-term ﬁxed costs associated with accepting a generator offer that requires a unit to start-up, is in a sense treated as a separate, discrete product with its own pricing rules (under the revenue sufﬁciency payment), as is operating the generator at the minimum or “no-load” level. A parallel construction is found on the demand side, in which the demand may bid to consume MWh along with short-term ﬁxed costs associated with implementing a demand curtailment offer. The pricing and settlement rules are discussed further in this section as well as extensively in Sections 5.3–5.6. The ancillary services are reliability services offered by eligible suppliers or responsive demand. As summarized in Table 5.1, these include regulation and different types of operating reserves, measured both as energy and as capacity (MW) made available at locations on the transmission network in one of the hourly markets (day-ahead or realtime). The spatial aspect of regulation and reserve product deﬁnitions is typically slightly different from that of energy. The ISOs deﬁne locations for these products on a zonal basis. These products share the characteristic that they are difﬁcult to disaggregate among each buyer on the system, so they are currently procured by the ISO on behalf of buyers, who pay a load-weighted average price. Because regulation and spinning reserves are complementary or substitute uses of generator providing energy (or a load consuming energy), their deﬁnition typically incorporates an energy component. The implications for auction design are discussed below.

EBL

184

Competitive Electricity Markets

Table 5.1. Deﬁnitions and characteristics of market-based ancillary services in the US ISOs currently in operation or under consideration Type of ancillary service

Description

US ISOs with markets

Regulation (or automatic generation control, AGC)

The ability to increase or decrease energy output on a second-by-second basis for energy balancing Reserves available (MW) within 10 minutes from generators synchronized with the grid or demand response Reserves available (MW) within 10 minutes from generators not synchronized with the grid or demand response Reserves available (MW) within 30 minutes or more from generators either synchronized or not synchronized with the grid or demand response A product of generators and types of transmission elements that essentially supports the voltages that must be controlled for reliability

New York ISO, ISO New England, PJM, California ISO

Ten-minute spinning (or synchronous) reserve

Ten-minute non-spinning or non-synchronous reserve Thirty-minute or supplemental reserves

Reactive power

New York ISO, ISO New England, PJM, California ISO

New York ISO, ISO New England, California ISO

New York ISO, ISO New England, California ISO

Currently cost-based procurement; markets under consideration (see, e.g., FERC, 2005).

EBL 5.2.2. Functions of the ISO A major market design question in the mid-1990s concerned the relationship between the transmission system operator and the operator of the spot markets. This issue takes on different forms in different countries, due to the requirements of their regulatory systems and the history of the electric industry. For example, in some countries (such as England and Wales), the transmission system operator is a single ﬁrm under performance-based regulation (see, e.g., Chapter 4). In the United States, there was no such opportunity to merge the ownership of national or regional transmission operations at the start of restructuring. Instead, there has been an ongoing experimentation on a regional basis with different organizational approaches to the implementation of transmission open access and management of transmission usage. And as a result, options for designing the energy and ancillary service markets also take different forms in different regions of the United States. As noted in Section 5.1, there are currently basically two organizational approaches to transmission system operations in the United States: individual utility operations and ISOs. To understand the purpose behind the ISO market, it is worth beginning with a description of markets where the grid is operated by individual utilities. In such markets, all energy transactions (long-term and short-term) are made by individual buyers and sellers (i.e., bilaterally) but there is no central trading point that co-ordinates transmission usage. Hence, before or after an energy contract is in hand, the buyer or seller must arrange a transmission contract for a pre-speciﬁed period that matches the energy contract’s source point(s) and sink point(s), and contains other characteristics such as the priority given to the transmission reservation (sometimes called the “ﬁrmness” of

Design of US Wholesale Energy and Ancillary Service Auction Markets

185

the transmission reservation).3 Ancillary services, including balancing energy, are bought under cost-based rates (FERC, 1996a). This type of market has supported an increase in wholesale energy transactions, especially with the advent of the Internet (FERC, 1996b). Its main limiting factors as a market design are the following: •

The lack of co-ordinated transmission operations to efﬁciently utilize transmission capacity across multiple utilities, especially when there is congestion (FERC, 1999b, 2002). • The lack of a centralized exchange or auction to facilitate spot trading of energy or ancillary services (FERC, 2002). A higher degree of market organization, as offered by an ISO, is thus generally desirable to support market efﬁciency and expand market scope. Turning ﬁrst to organization of transmission operations, market improvements can be achieved in several ways, with different implications for market efﬁciency. First, individual (transmission-owning) utilities can retain control of their transmission systems, but can establish an independent entity (sometimes called a “Day 1 ISO”) to facilitate information exchange and transmission capacity reservation by parties seeking to buy transmission. This type of organization can improve transmission scheduling to facilitate energy trading.4 Second, individual utilities can cede full scheduling control over their transmission systems to an ISO, which will simultaneously operate the entire transmission network collectively along with real-time energy and certain ancillary service markets. Whether the ISO also operates energy and ancillary service markets prior to real-time, such as day-ahead markets, has been another design decision. Whichever type of centralized transmission operations is chosen, an organized market for electricity can be developed along several different formats. If that market is going to be organized at a central location, then it is likely to be either an exchange or an auction. An exchange is a common format for commodity forward and futures trading. This is typically an “open” market in that buyers and sellers are known to each other. The most common pricing rule is the “bid-ask” method characteristic of futures exchanges. In an auction, a third party known as an auctioneer matches buyers’ and sellers’ following a pricing rule. Either an exchange or an auction could be used to organize spot electricity markets, but for reasons discussed next, an exchange is more compatible with pre-dayahead forward markets while an auction is best suited to the day-ahead and real-time spot markets operated by ISOs.

EBL

5.2.3. Choices in electricity auction design The electricity markets described in this chapter are all auction markets. To expand on the description given above, Table 5.2 summarizes some common auction designs (see, e.g., Klemperer, 1999; Krishna, 2002). In general, the choice among the designs depends 3 In the US context, ﬁrmness designates the physical priority of a point-to-point transmission contract in the event that reliability concerns, such as unmanaged congestion, prompt the system operator to scale back (“curtail”) transmission usage. Typically “non-ﬁrm” contracts, which are usually shorterterm, are curtailed before “ﬁrm” ones. In the United States, this arrangement prevails only outside the ISO markets, which use price-based congestion management rather than physical priorities. See also discussion in Chapter 4. 4 The Midwest ISO operated as such a transmission scheduler from 1997 to 2005, when it began operating a market with the design described in this chapter.

Table 5.2. Comparison of some alternative auction designs

SingleUnit Auctions

Auction type

Description

Revenue adequacy

Incentive compatibility

Open ascending price (“English”)

A one-sided auction in which the auctioneer raises the price of the unit being sold until there is one remaining bidder. The selling price is the winning bid. A one-sided auction in which the auctioneer lowers the price of the unit being sold until one bidder is willing to purchase. The selling price is the winning bid. A one-sided auction with sealed bids in which the selling price is the highest bid price.

Yes

Sealed-bid second price

A one-sided auction with sealed bids in which the highest bid wins the auction, but pays the bid of the next highest bid.

Yes

Yes. The bidder with the highest valuation would remains until all other stop bidding. The winner pays the valuation of the next highest bidder. The bidder with the highest valuation for the item would have the incentive to wait past where its value is in order to receive a lower price. No. The bidder with the highest valuation would want to shave its bid to receive a lower price and thus could lose out on winning the auction Yes. Each bidder would have the incentive to submit its own valuation.

Sealed-bid uniform price

A one- or two-sided auction with sealed bids in which the market clearing price for all units is the highest bid price for any unit. A one-sided auction with sealed bids in which each seller that clears the auction is paid its offer price for any individual unit.

Yes

Open descending price (“Dutch”)

Sealed-bid ﬁrst price

MultipleUnit Auctions

Sealed-bid discriminatory (pay-as-bid) Sealed-bid Vickrey

Open auctions (English or Dutch)

EBL

Yes

Yes

Yes

No

These are similar in design to the single-unit English and Dutch auctions.

Sources: Krishna, 2002; Hobbs et al., 2000.

Similar to single unit auctions.

Not completely. If the bidder can inﬂuence the price, it would have the incentive to buy less in order to lower the price for the amount it buys No, the bidder would have the incentive to guess what the clearing price would be. Yes, each bidder would pay the opportunity cost imposed on the other players if the bidder had not been in the market, and as such would have the incentive to bid its true valuation. Similar to single-unit auctions.

Design of US Wholesale Energy and Ancillary Service Auction Markets

187

on a number of factors, including the characteristics of the product being auctioned and the revenue and efﬁciency properties of each auction type. The typology in Table 5.2 divides auctions into either “single-unit,” in which buyers and sellers are trading single items or single bundles of items (such as a single MW or a “strip” of MW over multiple hours), or “multiple-unit,” in which multiple units are traded simultaneously.5 As noted, these types of auctions can then be further divided into open or sealed bid formats. Under the sealed bid format, buyers and/or sellers submit bids and offers that are not known to others. The anonymity of a sealed bid auction is further compounded in the spot electricity auction, in which buyers do not know which seller they are buying from. Within the multiple-unit, sealed-bid auction framework, the two major choices for pricing are discriminatory, or “pay-as-bid,” prices or uniform clearing prices. Because auction designs must ﬁt the commodity being traded, not all the auction designs shown in Table 5.2 are applicable or desirable for the electricity commodities discussed in this chapter. The product deﬁnitions and the technological characteristics of wholesale electricity vary in the different stages of the market: the pre-day-ahead forward markets that take place outside the ISO markets, where trading is primarily for multi-day/multihour strips of power; the day-ahead markets, where trading is on an hourly basis; and the real-time spot markets where power goes to physical production and delivery and trades takes place within the hour itself. The clearest way to describe how auctions can be designed around these forward and spot markets is to begin with the ISO’s real-time market and work backwards in time. The real-time market has a number of characteristics that greatly narrow the alternatives for auction design, at least with current power system technology. In real-time, the system operator is balancing supply and demand on a time-frame of seconds and minutes, largely by directing the output of generators on the system over such time-frames (and some aspects of the transmission facilities). Because of the nature of power ﬂows, the injections and withdrawals of power on the system must be balanced simultaneously; i.e., individual injection and withdrawal combinations cannot be evaluated independently. Hence, there are no physical “bilateral” transactions involving matching between buyers and sellers in real-time: all purchases are either via the ISO market or self-provided.6 The real-time dispatch of generators must also account for inter-temporal constraints, such as ramp rates, that create interdependencies across hours. Moreover, demand in real-time can be price-responsive, but typically not on a time-frame of minutes or seconds; hence, such demand will inevitably remain largely price-inelastic until technology provides greater responsiveness and the number of units of electricity (e.g., MW) consumed will ﬂuctuate from minute to minute. These factors inevitably require that a real-time auction is multiple-unit and sealed-bid and that the market be simultaneously cleared with all relevant transmission and generation constraints taken into account. That is, it is simply not possible to have a single-unit electricity auction in real-time, even for multi-hour strips (although that can be done in the forward markets) and it is hard to conceive of an open auction format that can meet the time-frames necessary. Essentially, real-time electricity auctions as they currently function in the United States are simply least-cost (“economic”) security constrained dispatch

EBL

5 Multiple-unit auctions do not have to take place simultaneously, but could instead take place in a sequence of single-unit auctions (Krishna, 2002); however, this latter format is not consistent with the physical characteristics of electricity. 6 Prior to the formation of ISOs, individual utilities provided other parties that were using their transmission facilities with “energy balancing,” typically priced on an average basis.

188

Competitive Electricity Markets

using market-based supply offers and demand bids to determine prices. These auctions will be described in more detail below. Whether the pricing rule should be uniform or discriminatory will also be discussed in the next section. In the forward, but short-term, electricity markets, more auction design alternatives become possible, including some simpliﬁcations of commodity pricing, although still constrained by their proximity to the physical requirements of the operating day. For close to real-time auctions, such as hour-ahead or day-ahead auctions, deviations from the auction design for the real-time market are possible, such as relaxing some of the physical constraints on the network, but consideration must be given to implications of such designs for reliability and economic efﬁciency. Economic efﬁciency will be adversely affected if the forward auction schedule – i.e., the scheduled output of generators going into the operating day – is substantially different from the real-time dispatch, and if the resulting adjustments for dispatch feasibility incur otherwise avoidable costs or if such costs are assigned on a basis that undermines efﬁcient price signals. For example, in the early phases of ISO auction design in the United States, some parties preferred day-ahead auctions or exchanges that did not meet the requirement of simultaneous feasibility of the resulting day-ahead schedule using an actual network. Instead, they argued for a zonal approximation of the network to facilitate day-ahead trading. This was the design choice adopted in the ﬁrst phase of the California market (1998–2001), in which there was a separate day-ahead market operated by the California Power Exchange (PX) and a real-time market operated by the California ISO. To facilitate the PX market, California was divided into two transmission zones that largely, but not exactly, corresponded to the major transmission constraints on the system. The PX operated a single market in each zone using a multi-unit auction with a single zonal clearing price (see, e.g., Sweeney, 2006). The resulting schedule was then passed to the ISO, which could aim to adjust it to the limits on inter-zonal transmission through adjustments offers and bids available from the PX, but which could not enforce an efﬁcient schedule by altering the scheduled output of generators on the basis of their supply offers. In real-time, the ISO would have to make any ﬁnal scheduling adjustments on both an inter-zonal and an intra-zonal basis using real-time energy and regulation offers. In concept, the PX was to be joined over time by other competing day-ahead exchanges which could all operate their own auction markets or exchanges under whatever design suited them. The ISO would thus have been collecting many day-ahead schedules and attempting to impose feasibility on them in a similar fashion. Over time, evidence collected that, along with creating cross-subsidies on an intra-zonal basis, such zonal pricing did not provide appropriate locational signals for investment in generation, transmission, and demand response. California ISO will shortly revise the market design to adopt the day-ahead and real-time auction with locational marginal pricing described in this chapter. Like energy used for balancing, the real-time markets for regulation and operating reserves are most compatible with the sealed-bid, multi-unit auction design with all network constraints represented and uniform prices. Again, day-ahead markets for these services could use other auction designs but, as with energy, potentially resulting in market inefﬁciency. For these products, the primary design issues have been (i) the sequence in which the energy, regulation, and reserves markets are cleared; (ii) whether they are co-optimized or not; (iii) how complementarities and substitutions between them are captured in the auction pricing algorithm; and (iv) the components of the supply offers. These design issues have been reviewed in FERC (1999a, 2002), Stoft (2002), and O’Neill

EBL

Design of US Wholesale Energy and Ancillary Service Auction Markets

189

et al. (2006), among others. The US ISOs that offer these ancillary services through auctions have largely converged on a design that incorporates co-optimization and represents a hierarchical substitution that supports an efﬁcient use of generators and any demand resources that can provide them. More detail on these auctions follows in Sections 5.3 and 5.5. 5.2.3.1. Uniform versus discriminatory (pay-as-bid) pricing As noted, within the multiple-unit, sealed-bid auction format, there are two primary pricing alternatives: uniform or discriminatory, also known as pay-as-bid (Klemperer, 1999; Krishna, 2002). At present, there is no theoretical consensus on the revenue and incentive compatibility properties of these rules as applied to electricity auctions (see, e.g., Kahn et al., 2001; Fabra et al., 2004). However, there is agreement on their implications for market transparency and also market power mitigation, as described later. Those implications have made the uniform pricing rule more attractive in the US ISO markets (although the pay-as-bid rule is currently used in the England and Wales spot market). There are also variations within each type of pricing rule that have implications for market efﬁciency. Under uniform pricing, all sellers are paid the price offered by either the last unit (MWh) chosen by the auction or the next (incremental) unit that is not chosen (the exception to this rule in some ISOs is in periods of scarcity pricing, as discussed shortly, during which the market price is set through a demand curve). For example, if a $40/MWh offer is accepted and the next offer in the offer stack that is not accepted is for $50/MWh, then the clearing price for all sales is set at either $40/MWh or $50/MWh. In the real-time market for energy, the choice between these alternative uniform pricing rules has to do with how the ISO seeks to control the output of the marginal generator(s) through a combination of price signals and quantity instructions. The choice also has an impact on long-term pricing signals. For example, if due to generation constraints, such as generator ramp rates (and assuming no transmission congestion), the ISO has to turn on a $50/MWh unit while a $40/MWh unit is operating at below its economic maximum output, but that the next available MWh is at the $40 price, then the choice of pricing rules has clear incentive properties:

EBL

•

If the market price is set at $50/MWh, then the $40/MW unit may have an incentive to increase its output. If this creates reliability problems for the system operator to manage, then it will have to control the output of the $40/MWh unit through dispatch instructions and penalties for deviations (as is done in many ISOs). • If the market price is set at $40/MWh, then the $40/MWh unit has no incentive to increase output (assuming that its offer is reﬂective of its marginal cost), but an alternative pricing rule has to be established to pay the $50/MWh unit to turn on. Typically, the unit is paid its bid and the ISO collects the revenues from buyers. On an electrical network with nodal congestion pricing, this auction pricing rule results in uniform clearing prices at each location and is called locational marginal pricing. A more detailed example of this pricing rule is provided in the numerical example that begins in Section 5.4. In pay-as-bid designs, the supply offer is typically conceived as a bundled offer for energy, start-up, and no-load (sometimes called a “one-part” offer). For example, if a supplier makes an offer of $20/MWh and is cleared through the auction, then the offer

190

Competitive Electricity Markets

is paid $20/MWh, no more or less. Under most proposed formats, buyers in the auction would pay an averaged price based on the total accepted MWh. Hence, while a locational pricing result could be sustained for sellers, it would be difﬁcult to sustain for buyers, except on an aggregate basis, such as in a zone. Advocates of pay-as-bid pricing have different motivations, depending on whether they are sellers or buyers. Sellers may believe that the combination of one-part offers with payas-bid provides the seller greater transparency over the auction result than markets with three-part offers and locational marginal pricing and additional payments to guarantee recovery of start-up and no-load offer costs. Buyers may believe that a pay-as-bid rule will result in a lower market clearing price. This belief is sometimes based on the perception that in a competitive pay-as-bid auction, a unit with a $20/MWh marginal cost will always offer at that price, even if other units are clearing at a higher price. This argument is incorrect. As long as there is sufﬁcient market price transparency, then a supplier will always seek to obtain the price that clears the market to maximize proﬁt. Hence, in a pay-as-bid market, the unit with a $20/MWh marginal cost will raise its offer price to its estimate of the price offer of the most expensive unit cleared in the same auction time period. This incentive raises the following concerns with pay-as-bid auctions. First, the need for sellers to estimate the hourly market clearing price could lead to many incorrect guesses, which even in a competitive market will lead to inefﬁciency (e.g., when a lower-cost unit overestimates the price, letting a higher-cost unit clear the market). Second, from the perspective of the market operator, efﬁcient control of the power system will become less transparent as all offers for a particular hour will cluster around the market clearing price. There is the argument that a low-price unit will be risk-averse in a pay-as-bid auction and rather than attempt to clear at the estimated market price will persistently shave its offer to ensure that it is scheduled. That is, a $20/MWh unit will offer at $50/MWh rather than at $60/MWh, which is the price of the most expensive unit chosen. Such risk aversion would lower the price to buyers. However, over time, it will also send the wrong price signal for investment, resulting in higher-cost plants being built. Hence, a short-term lowering of market prices may result in an increase in prices over the long term (Cramton and Stoft, 2006). Finally, because all offers will likely cluster around the estimate of the market-clearing price in each auction period, if there is the need for ex ante or ex post market power monitoring and mitigation, the regulator will have more difﬁculty reconstructing which units have been raising their offers above marginal cost in an attempt to manipulate price levels. This is discussed later in this section and in Section 5.9. For all these reasons, US regulators have preferred the uniform clearing price rule to the pay-as-bid rule. However, in at least one other country (England and Wales), regulators have actually redesigned the spot market to move away from uniform pricing rules, adopting pay-as-bid pricing.

EBL

5.2.4. Sequencing of markets and reliability functions As noted earlier, the basic design choices for real-time electricity auctions are highly constrained by the characteristics of power system operations. Reliability requirements, timelines for operational decisions in the hours prior to real-time, and representation of generation and transmission capacity constraints all further shape the forward market designs by adding mathematical constraints to the auction solution and requiring speciﬁc

Design of US Wholesale Energy and Ancillary Service Auction Markets

191

market and “out-of-market” procedures. Over the decade of ISO auction market development, two major questions stand out in this regard: •

If there are forward energy or ancillary service auctions conducted by an ISO or any entity able to offer into an ISO market, how far forward in time should system reliability and operational constraints be applied? Put another way, how should the interface between forward and real-time markets be designed? • How should the physical constraints of power system operations be reﬂected in the ISO day-ahead and real-time auction pricing rules? Both of these questions were central in the market design discussions that took place in the United States in the mid-1990s (see, e.g., Stoft, 2002; Wilson, 2002; O’Neill et al., 2006). With regard to the ﬁrst question, an initial debate placed those who believed that system operations could be a primarily real-time function, with forward markets operating separately under their own rules until just a few hours or less before real-time, against those who argued for a full integration of short-term (day-ahead and real-time) market and system operations. The second question concerned issues such as the choice of zonal pricing versus locational marginal pricing in transmission usage pricing and whether supply offers into the auctions should give generators the choice to represent details of their short-term marginal costs, such as start-up costs. The FERC standard market design proposal (FERC, 2002) sought to settle as many of these design debates as possible. As it proposed, and is now done in most of the ISO markets, the sequence of short-term auction markets and reliability actions incorporates all relevant reliability procedures and physical constraints and provides sufﬁcient offer and bid detail to provide an efﬁcient auction result. As the remainder of the chapter will explain, this sequence has the following structure: •

EBL

First, the ISO undertakes a series of pre-day-ahead procedures and out-of-market actions to account for changes in system conditions as well as demand (load) forecasting and scheduling of generators with longer than one-day start-up or shutdown requirements. • Second, the ISO operates a day-ahead market that includes an auction with “securityconstrained unit commitment,” i.e., consideration of the full set of known transmission security and generation unit constraints, within the limitations of the auction optimization. • Third, the ISO takes several types of additional actions after the day-ahead market clears to ensure reliability prior to real-time. Most notably, all ISOs undertake variations on what can be called reliability unit commitment, commitments of additional generation to replace physical units displaced by virtual supply and to meet forecasts of actual load if such forecasts are different from the amount of demand that is cleared day-ahead. In addition, the ISO collects data on generation cleared day-ahead to determine changes in actual availability. • Finally, in the real-time market, as discussed earlier, the ISO operates the power system through a security constrained economic dispatch (complemented by unit commitments) using offers and bids to determine auction market prices and supported also by physical dispatch instructions that may occasionally differ from the market result for reliability reasons. 5.2.5. Market power monitoring and mitigation Market power in the electricity auction markets is deﬁned as the ability of a buyer or seller to signiﬁcantly and sustainably alter the market price from the competitive price. Most

192

Competitive Electricity Markets

economists interpret the competitive price as the market price that results when sellers are willing to offer power at their marginal opportunity cost and buyers bid for power at their true willingness to pay. The general economic principle is that the larger the number of sellers and buyers, the more likely that market prices and quantities will be competitive. This principle is measured through indices of market concentration or by market price simulations.7 Since there are currently few price-responsive buyers in electricity spot markets, the auction price is usually set by sellers (except when administrative scarcity pricing is enforced). Hence, the competitive price is typically estimated or simulated based on known production costs of the marginal unit delivering to a location, primarily fuel costs, and, if possible, adjusted to account for short-term ﬁxed costs (such as start-up) and inter-temporal opportunity costs (for limited energy plants, such as hydro or emissions-constrained facilities). When the market price is above this competitive price, which it usually is, then some degree of supplier market power is being exerted. In economic terms, consumer surplus is being transferred to producers and the total producer surplus is being shifted among suppliers. The task of the regulator is to determine whether and how to manage such market power such that a reasonable approximation of competitive market prices and quantities prevails. The legal and regulatory methods for doing so vary between countries and supra-national organizations, such as the European Union. In the United States, FERC has the statutory obligation and authority to mitigate the market power of sellers and buyers in the wholesale electricity markets under its jurisdiction.8 Although there have been occasional concerns with buyer market power, the primary regulatory concern in the United States has been with seller market power.9 In general, there are essentially four types of measures through which seller market power in an ISO auction market (or other electricity markets) could be mitigated. First, ex ante structural measures can be taken to enhance the competitiveness of the auction market prior to the market start. For example, if regulators ﬁnd that the market is too concentrated, ownership of generation assets can be restructured, through divestiture, to diminish the market power of sellers. In the United States, FERC has not required any utility to divest generation prior to selling wholesale power; instead, FERC can selectively prevent suppliers with excessive market power from selling wholesale power at market prices by approving sales only at regulated cost-based rates.10 In practice, this procedure has not

EBL

7

Most ISOs now include such measurements of market power in their annual state of the market reports. 8 In the United States, the Federal Power Act requires that generation and transmission market prices are “just and reasonable.” This standard has been interpreted by the courts as giving FERC the authority to monitor and mitigate market power in the electricity markets. See discussion in O’Neill et al. (2006) and Helman (2006). 9 With respect to buyer market power, a notable design decision was taken in the ﬁrst phase of the California wholesale market design to require the incumbent vertically integrated utilities to purchase all their wholesale power through the daily markets in part to suppress their buyer market power. See, e.g., discussion in Sweeney (2006), p. 338. 10 FERC screens prospective sellers, individually or collectively (e.g., as market participants in an ISO), for generation market power using various measures of market share and market concentration. Only sellers that pass the screens are automatically granted the right to sell at market prices (called “market-based rates”); those that fail must either prove their lack of market power using additional information or sell at a regulated cost-based rate or some other negotiated rate. See, e.g., Helman (2006).

Design of US Wholesale Energy and Ancillary Service Auction Markets

193

been used to screen individual market participants in ISO markets. Rather, sellers are allowed to participate in the ISO auctions subject to the ISO’s market power mitigation rules. The ISO market power rules are predicated on the assumption that while the market is structurally competitive in the aggregate, some sellers have market power on an intermittent basis, such as when transmission constraints subdivide the market or during peak hours when most or all generators must run. This has led to a second type of market power mitigation: the application of ex ante behavioral measures. These usually take the form of offer caps that limit the price that a supplier can offer into the auction. The ISO auction markets have relied primarily on such offer caps, as discussed in Section 5.8. These caps can apply at all times to all bidders (as the $1000/MWh offer caps do in the eastern US ISOs), or they can be selectively imposed if there is evidence that offer prices are raised signiﬁcantly above competitive levels and also would affect market prices. Third, the regulator and the ISOs can apply ex post measures, such as penalties for certain types of behavior or refunds to buyers if the auction market prices are found to be in excess of competitive levels, within some margin of error. FERC has the legal authority to apply such ex post measures, as it did, e.g., in requiring refunds following the California market crisis (see, e.g., Sweeney, 2006). It is a policy decision how to strike a balance between ex post and ex ante measures. In general, ex ante measures provide for less market uncertainty than ex post measures. For example, offer caps can be seamlessly integrated into the auction pricing rules, as discussed in Section 5.8. Finally, there are auction designs that mitigate market power without the need for structural or behavioral measures. These are called Vickrey–Clarke–Grove mechanisms, and they provide incentives for truthful bidding and, thus, efﬁcient operation (Hobbs et al., 2000). Such an auction design could in theory be applied to electricity markets, but poses a number of practical issues. Perhaps the most serious issue is that such a mechanism would not be revenue neutral; in general, the auctioneer would pay more to suppliers (in a sense, to pay off their market power) than they would receive from market buyers. Hence, the preference of ISO market designers has been to retain the auction design with sealed bids and uniform price clearing, but apply various methods of ex ante and ex post market power mitigation.

EBL

5.2.6. Scarcity pricing The advent of supply offer caps for purposes of market power mitigation and the lack of demand bids has created a new design issue for the electricity auctions, namely that spot market prices cannot always reach levels sufﬁcient to reﬂect the true scarcity of supply or the true willingness-to-pay of demand. In markets without such economic regulation or market incompleteness, there is no distinct feature called “scarcity pricing.” Rather, prices rise and fall largely in response to the supply–demand balance. Under scarcity conditions, supply is short and prices are high relative to historical norms, so consumers voluntarily or involuntarily reduce their consumption of the commodity in question. In the US electricity markets, empirical evidence emerged that some existing generators in ISO markets were not revenue sufﬁcient due to offer caps (and were seeking to return to regulated status under types of cost-based contracts) and that investment in both generation and demandresponse capability was not adequate. Largely because the electricity markets are not considered sufﬁciently competitive in all market conditions to remove or greatly relax the

194

Competitive Electricity Markets

offer caps, FERC and market operators have turned recently to administrative scarcity pricing as a means to complete the market design.11 As of this writing, various scarcity pricing rules are being implemented, and some proposed, in the ISO markets. The ﬁrst design issue is to deﬁne a scarcity condition in the market. In the current electricity markets, with largely price inelastic short-term demand, scarcity is clearly indicated when demand is either voluntarily or involuntarily curtailed. From a system operational perspective, operating reserve shortages are the ﬁrst indicator to the market operator that demand curtailments are more likely. Hence, most ISOs are using, or considering using, operating reserve shortages as a trigger for scarcity pricing. The second design issue is to determine what the scarcity price should be and how it should be set. The general proposed approach is to raise the price above the highest offer cap incrementally, corresponding to the degree of reserve shortage or demand curtailment. One approach is to create a “demand curve” that begins to automatically set the price as soon as reserves are short.12 The market price is determined by the intersection of the available reserve quantity and the demand curve, with the highest price being reached when reserves are at a minimum allowable level. A simpler method is to raise the price to a single high pricing point when reserves are short to create an adder to the energy market price. Most analysts would prefer that the scarcity pricing curve be related to a measure of the value of lost load (VOLL) (see, e.g., Stoft, 2002, 154–64). However, there is no consensus on how to determine VOLL exactly, and high VOLL prices may be politically unacceptable. Hence, as will be discussed further below, while different ISOs have reached their own conclusions, there is as yet no consensus on the parameterization of scarcity pricing curves. With this background on how auction design decisions have been made and how regulatory, technological and administrative factors have motivated modiﬁcations to the auction designs and pricing rules, the chapter turns next to the details of the spot electricity auctions, beginning with the day-ahead market.

EBL

5.3. The Day-Ahead Market The day-ahead market begins the sequence of short-term market-driven and operational procedures that lead to the daily efﬁcient and reliable functioning of the regional power system under ISO control. The day-ahead market encompasses a day-ahead auction market for energy as well as regulation and operating reserves that follows the basic auction format described in Section 5.2, along with many other details that will be described here. The day-ahead market also includes non-price schedules that are included as constraints in the auction market clearing. These schedules include generation and load self-schedules, bilateral schedules, and import and export schedules. As currently designed, the day-ahead market is best described as a forward market subject to all the physical and reliability power system constraints that are known at the time to affect the next-day (real-time) dispatch. It is a forward market because sales or purchases cleared at the day-ahead price that are not subsequently converted into a physical position in real-time must be “sold back” or “bought back” at the real-time market 11

The initial expectation in most ISO markets was that the $1000/MWh offer cap would to allow suppliers to set a sufﬁcient scarcity price using their offers (see, e.g., FERC, 2002). 12 Under scarcity pricing the price is set by the administratively determined demand function.

Design of US Wholesale Energy and Ancillary Service Auction Markets

195

price.13 This design principle generally applies to any generation product sold in the dayahead market, whether energy, regulation, or operating reserve (although as discussed below, in practice ISOs may not fully implement this principle for reserves). It also applies to energy purchased in the day-ahead market, but not to regulation or reserves (which the ISO procures on behalf of buyers). This sequence of ﬁnancial settlement generally creates an inducement to transact in the day-ahead market based on a forecast of the next-day supply–demand balance. The “physical” aspect of the day-ahead market is that the auction ﬁnancial offers, along with any submitted schedules, are subject to a number of physical constraints on market clearing, including generator constraints, such as ramp rates and minimum and maximum output levels, and transmission network constraints. As noted above, With the inclusion of offers to start-up generation units, this is called security-constrained unit commitment. The initial objective of market designers was that security-constrained unit commitment, in concert with the ﬁnancial incentives noted above, would shape the day-ahead schedule into a reasonable approximation of a feasible real-time dispatch. This would give the ISO power system operators time to evaluate system conditions in the spot markets, under which suppliers are no longer necessarily dedicated to particular buyers, before the operating day began. However, with the introduction of “virtual” sellers and buyers (deﬁned in the next section), which can displace physical offers and bids, the day-ahead schedule has become less physical and more ﬁnancial. To address this, ISOs introduced additional reliability unit commitments following the day-ahead market, as described in Section 5.4. Hence, the day-ahead market, taking into account all relevant system constraints, in concert with the reliability unit commitment allows the ISO to reduce scheduling uncertainty about the next-day dispatch.

EBL

5.3.1. Market procedures The day-ahead market trading rules are fairly consistent across the ISOs. Offers and bids in the day-ahead market, along with non-price schedules, are due typically by midday of the prior day, although as shown in Table 5.3 some ISO markets open and close earlier in the day. Offers and bids for energy and ancillary services are divided into price offer components and physical parameters. These are summarized for two ISO markets, PJM and New York ISO, in Table 5.7. As noted already, in each day-ahead ISO market there are three price components for energy suppliers, which also support the provision of regulation and reserves: start-up ($), no-load ($/MWh), and energy, sometimes called “incremental” energy ($/MWh). The energy offer represents the seller’s minimum dollar value that it is willing to accept to supply energy. A negative energy offer represents the seller’s maximum willingness to pay to produce power at a particular level of output (usually the minimum physical load level of the generation unit). The markets for regulation and operating reserve may include an additional offer component ($/MW) to represent costs associated with providing those services. The remaining generation offer components are physical parameters, which include upper and lower operating limits, both under normal and emergency conditions, ramp rates, minimum run times, maximum starts per day, and other parameters. Demand bids for energy may also have multiple components. In general, the core feature of demand bid is a MW block with a $/MWh reﬂecting the buyer’s 13

In US regulatory jargon, this is sometimes called a “two-settlement system,” referring to the dayahead (ﬁrst) settlement and the real-time (second) settlement.

Competitive Electricity Markets

196

Table 5.3. Scheduling, pricing, and settlement timelines for day-ahead and real-time energy in the US ISOs PJM

New York

New England

MISO

Day-ahead market offer period closes (prior day) Day-ahead market results posted (prior day) Reliability unit commitment opens for offers (prior day) Reliability unit commitment offer period closes (prior day) Reliability unit commitment results posted (prior day) Real-time market opens for offers Real-time market offer period closes

12:00

05:00

12:00

11:00

16:00

11:00

16:00

16:00

16:00

17:00

18:00

Integrated with day-ahead market 05:00

18:00

18:00

by unit

11:00

by unit

20:00

Real-time market preliminary results posted

During hour (Integrated 5 minute LMPs)

16:00 (prior day) 18:00 (prior day)

75 minutes prior to the (dispatch) hour 30 minutes prior to the (dispatch) hour

EBL

16:00 (prior day) 18:00 (prior day) During hour (Integrated 5 minute LMPs)

30 minutes prior to the (dispatch) hour During dispatch hour (integrated 5 minute LMPs)

Sources: ISO and RTO technical manuals (see references).

maximum willingness to pay to consume power. Demand bids may also submit ﬁxed cost components, such as an additional cost to shut down a particular piece of equipment. As Table 5.7 illustrates, ISO vary in the periodicity of day-ahead offers and bids and the frequency with which they can be changed. The implications of these differences will be discussed below. Beginning in 2000, the eastern US ISOs began to introduce virtual supply offers and demand bids into the day-ahead markets. These are purely ﬁnancial positions in the forward market that will not be converted into a physical position and must be re-settled at the real-time market price. An accepted virtual supply offer that “sells” energy at a day-ahead LMP has to “buy back” that energy at the real-time price at the same location. Similarly, an accepted virtual demand bid that “buys” energy day-ahead has to “sell” it back in real-time. There are a number of purposes for such virtual transactions. A common use by purely ﬁnancial entities (i.e., those that have no physical positions) is to arbitrage the price spread between the day-ahead and real-time markets. Hence, a virtual seller seeks to sell day-ahead energy at a higher price than it will have to buy the energy back in realtime; a virtual buyer has the opposite objective (see the numerical examples in Sections 5.3 and 5.5). With entry into the virtual trading market, competition between virtual traders for arbitrage rents causes greater price convergence between day-ahead and real-time.

Design of US Wholesale Energy and Ancillary Service Auction Markets

197

There are other uses for virtual transactions. For example, if an entity holds ﬁnancial transmission rights, it can use such virtual positions to shift settlement of the transmission rights from day-ahead to real-time, if that is seen as ﬁnancially advantageous.14

5.3.2. Market pricing and settlement The day-ahead auctions basically follow the uniform-price pricing rule discussed in Section 5.2.3 for energy, but with additional features to pay other offer components and with many other pricing rules. Prices are determined as follows: The auction is conducted by inserting the supply offers and demand bids into a dynamic optimization program that minimizes the cost of meeting demand (or maximizes social welfare, measured as the sum of consumer and producer surplus) for energy and ancillary services in each hour of the operating day subject to generation and transmission constraints. The generator constraints are represented in the auction model both through discrete, or integer, variables, such as the start-up decision, and continuous variables, such as the generator energy offer (represented as a stepwise or piecewise linear function over the range of output). The dayahead auction result is an hourly schedule of LMPs and non-binding dispatch instructions for each generator, indicating its output level and respecting its start-up costs, energy offers, ramp rates, and other constraints on output, such as minimum and maximum operating limits. The market price for energy is calculated by the auction algorithm as the shadow price associated with the energy balance constraint for each node (typically a bus on the network) in the high-voltage transmission network (see Appendix 5A). That is, LMPs can be calculated for each network location where energy is injected or withdrawn and also for transshipment nodes. ISOs will typically limit the number of LMPs calculated to nodes that have commercial purposes. The LMP is a composite of the accepted offer prices of all generators that would supply the next, or incremental, MW at that location. ISOs typically disaggregate LMPs into three components: energy, losses, and congestion (Schweppe et al., 1988). For computational reasons discussed in Section 5.7, the energy component is the same for all buses within an ISO’s network, while the loss and congestion components differ if there are losses and congestion. The ISO will generally collect surplus revenues when there are losses and congestion. Because the ISO is revenue neutral, these surpluses are returned to market participants through ﬁnancial transmission rights and loss refunds. One reason for disaggregating the LMP is that the ﬁnancial transmission rights issued by ISOs typically only cover differences in the congestion component, and not the loss component. Hence, the ISOs

EBL

14 A ﬁnancial transmission right as currently speciﬁed in the ISO markets is a contract to receive the difference in price between two locations on the transmission network for a particular quantity (MW) at a particular time. Financial transmission rights are cashed out in the day-ahead market. An entity holding a ﬁnancial transmission right can take an equivalent virtual position to create the exact opposite of the locations and MW speciﬁed in the right for settlement purposes. That is, it would bid to buy equivalent MW at the location where it was to withdraw the MW speciﬁed in the transmission right and offer to sell at the location where it was to inject the MW speciﬁed in the transmission right. This would exactly cancel its ﬁnancial transmission right position day-ahead, but would require it to cash out the equivalent position in real-time due to the virtuals. So the result would be that it buys back at the injection location and sells back at the withdrawal location at their respective real-time prices.

198

Competitive Electricity Markets

need to calculate how much surplus is generated by each component. There are various ways to calculate these components and the corresponding auction surpluses, which will be discussed in the numerical example that begins in this section as well as in Section 5.7. While the day-ahead LMPs will always provide a generator accepted through the auction with sufﬁcient revenues to cover its energy offer price over the hours that it operates, those prices may not cover its start-up and no-load costs. The numerical example that begins in this section illustrates this possible outcome. The ISO ensures that such offer costs will be recovered through a revenue sufﬁciency guarantee. This is an additional payment that takes place at the end of the day-ahead market, the real-time market, or both markets. This payment is discussed in more detail in Section 5.6. In general, all generators that provide a price offer for energy and all price-responsive demand bids that are “dispatchable” and all virtual offers and bids are eligible to set dayahead LMPs. A dispatchable generator or demand resource is one that the ISO can dispatch up or down based on its offered supply function. In addition, some ISOs let congestion bids (i.e., offers for a maximum price difference between two nodes) set locational energy prices. However, dispatchable generators in some situations are not allowed to set the price. There is usually a straightforward operational and/or economic incentive reason for such a restriction. For example, generators that are scheduled at their minimum load level for some hours are not eligible to set the locational price. The reason for this rule has to do with incentives: if the generator could set the price in the hours scheduled for minimum output, it may try to produce energy to increase its revenues. This would undermine the ISO’s attempt to establish an efﬁcient schedule that respects inter-temporal constraints. In both the day-ahead and the real-time markets, most ISOs have instituted various ex post aggregations of LMPs for settlement purposes. One such aggregation is called “hub” pricing, in which a location-weighted price is calculated for a set of nodes where spot energy is sourced. Another such aggregation is “zonal” locational pricing for demand, in which a load-weighted average price is calculated for the territory of particular utilities that have retail customers. In some cases, such zonal pricing is used to settle the purchases of multiple utilities, thus embodying some level of cross-subsidy. Some of these zonal pricing subsidies reﬂect regional agreements among utilities and state regulators. However, such zonal pricing also inhibits the development of demand response, as the actual nodal price is not known by buyers and higher prices will be hidden by the averaging process. Transmission usage charges apply to any market participant that has withdrawals in the market. This pertains both to spot transactions and to non-price schedules. With respect to the latter, a buyer with a bilateral contract or a utility that remains vertically integrated and desires to operate its own generators has no requirement to purchase energy through the day-ahead (or real-time) market. However, there is a requirement by such schedulers to pay for marginal transmission usage, as measured by differences in spot LMPs between points of injection and withdrawal. As discussed in more detail in Section 5.7, the congestion component of such usage charges for any particular MW schedule are hedged by ﬁnancial transmission rights between those points for the equivalent MW schedule, while any surplus marginal loss charge payments are refunded by the ISO on some basis independent of particular schedules.

EBL

5.3.3. Markets for regulation and reserves As noted in Section 5.2.3, some day-ahead markets now incorporate co-optimized regulation and operating reserves markets, although others have only real-time markets for these

Design of US Wholesale Energy and Ancillary Service Auction Markets

199

products. Details on this procedure are provided in Section 5.3.7.15 The design principles for day-ahead auctions of these products generally follow the sequence of ﬁnancial settlement obligations described earlier. Eligible generators and responsive demand make offers for these services into the day-ahead market. ISOs also allow “self-supply” of regulation and operating reserves, whether from a utility’s own resources or via contract. In general, to ensure reliability, the ISOs require that these schedules are offered through the auction at a zero price. Unlike energy, buyers do not make speciﬁc bids for these products into these markets, which are procured on their behalf by the ISO. The quantities that clear in the day-ahead market are then transferred to the real-time market and deviations are priced at real-time prices.

5.3.4. Scarcity pricing As discussed in Section 5.2.6, ISOs are experimenting with alternative types of scarcity pricing that set market prices during operating reserve shortages. In general, the demand curves or price adders/caps used to determine scarcity pricing in the day-ahead market should be the same as those used in the real-time market, to avoid providing incentives to market participants to alter behavior in one market so as to affect the prices in the other. However, there is a practical difference in the triggering mechanism. Day-ahead, an administrative scarcity price will be determined instantaneously as part of a day-ahead market solution. In real-time, as discussed below, system operators may be taking nonmarket steps to maintain reliability, and hence the declaration of an operating reserve shortage will be subject to manual decision-making within the operating hour. Some speciﬁc scarcity pricing rules are discussed below.

EBL

5.3.5. Congestion management Because congestion management is a slightly different procedure in the day-ahead market than in real-time, the day-ahead procedure is described brieﬂy here. Essentially, congestion in ISOs with LMP is managed day-ahead “implicitly” through the auction market optimization. Put simply, the day-ahead market result is a schedule for each hour of the day that includes the effects of congestion and in which the congestion “charge” between any two nodes or buses on the system is the difference between the LMPs at those locations. In contrast, in real-time, congestion management is an ongoing process that requires the system operator to make operational adjustments from minute to minute and which may involve non-market decisions. That procedure will be discussed in Section 5.6.

5.3.6. Numerical example This section begins a numerical example of market pricing and settlements that will continue through several sections of the chapter to illustrate the different stages of the ISO energy auctions. The example is developed in the simplest fashion possible to demonstrate realistic characteristics of an auction market with locational marginal pricing, while 15

Co-optimization is less cumbersome day-ahead than in real time, because the real-time procedure requires constant updating of generator set points and available regulation or reserve capacity. Realtime co-optimization is described in Section 5.5.7.

Competitive Electricity Markets

200

allowing the reader to replicate and track auction results. The mathematical details of this example are provided in Appendix 5A. The example assumes a transmission network with three transmission lines connecting three buses, or nodes, as shown in Fig. 5.1. A bus is represented by a thick vertical line while a transmission line is the thin line or lines connecting the buses. The network has characteristics that allow for the calculation of both the marginal congestion and the marginal loss components of LMPs. The transmission line connecting buses 1 and 2 has a capacity limit of 350 MW in both directions (1→2, 2→1); the other lines do not have capacity limits that will affect the examples. To simplify the presentation of power ﬂows on this network, a DC approximation of the AC load ﬂow is used, in a version with quadratic line losses (Schweppe et al., 1988). This means that for any 1 MW injected at one bus and withdrawn at another bus, the percentage of that 1 MW that is withdrawn is a decreasing, non-linear (quadratic) function of the total MW ﬂowing on the line (the line “loadings”). As noted, all the US ISOs calculate the loss component of LMPs, so while adding losses to the example makes it somewhat less intuitive, it is reﬂective of how the actual prices are calculated. The power ﬂow equations that are being used in these examples are shown in Appendix 5A. All the demand in this example is located at bus 3; this simpliﬁes the presentation, as all power will ﬂow to this location in each example. The 24 hours of the day-ahead market are compressed here into three demand blocks: an off-peak demand of 950 MW, an intermediate demand of 1300 MW, and a peak demand of 1600 MW. To further simplify, these demands are price-inelastic. Virtual demand bids are not considered, but would be represented as price-sensitive demand blocks (as would any price-sensitive physical demand). There are ﬁve potential suppliers, whose parameters and ﬁnancial offers are found in Table 5.4. The location of the suppliers is shown in Fig. 5.1; physical generators are represented as the circles connected by a solid line to each bus, while the one virtual supplier is represented as a circle connected by a dashed line. A cheap “base-load” generator “A” is located at bus 1. The offer price of this generator is $15.00/MWh and its capacity is 1500 MW. A more expensive “intermediate” load generator “B” is located at bus 2. The offer price of this generator is $20.00/MWh and its capacity is 250 MW. Both of these generators have start-up costs. The base-load generator has a very high start-up cost but because it is assumed to be operating for most hours of the year, and hence is already running in hour 1 of the day-ahead market, the auction algorithm does not need to consider its start-up

EBL

Bus 1

Bus 3

D E

A

C Demand

B Bus 2 Fig. 5.1. Three bus, ﬁve supplier electrical network.

Design of US Wholesale Energy and Ancillary Service Auction Markets

201

Table 5.4. Supply parameters and offer prices

Generator Generator Generator Generator Virtual E

A B C D

Bus

Capacity (MW)

Energy offer ($/MWh)

Start-up price ($)

1 2 3 3 3

1500 250 200 300 200

15 20 40 50 41

Not applicable 1000 2000 100 0

Table 5.5. Results of day-ahead market simulations Scenario

Off-peak Intermediate Peak

Demand (MWh)

950 1300 1600

Supply (MWh) A

B

1018 1227 1288

190 250

Price at bus ($/MWh) C

D

E

1

15.00 15.00 197 15.00

2

3

16.13 20.00 54.49

17.31 20.36 41.00

cost. The intermediate generator has a start-up cost of $1000. At node 3, there are two peaking generators, “C” and “D,” with similar capacity, but different offer prices, as well as one virtual offer, “E.” Peaking unit C has an energy offer price of $40.00/MWh and a start-up price of $2000; peaking unit D has a higher energy offer price of $50.00/MWh, but a lower start-up price of $100. The virtual offer E is priced at $41.00/MWh to exploit the price gap between the peaking units. Virtual supply offers could also exploit other jumps in the supply function in this example. The results of each demand scenario are summarized in Table 5.5 and in the Figs 5.2a–c. Beginning with the off-peak demand scenario depicted in Fig. 5.2a, the base-load generator at node 1 can supply all the power needed to meet demand of 950 MWh. The transmission capacity constraint on line 1 ↔ 2 does not bind, i.e., does not create congestion. However, the quadratic line losses require the base-load generator to inject 1018 MWh to meet the load (meaning that approximately 68 MWh are lost due to line losses) and the losses also establish differences in the LMPs. The prices at buses 1, 2, and 3 are $15.00/MWh, $16.13/MWh, and $17.31/MWh, respectively. At the LMPs for the off-peak scenario, the ISO auction charges demand $17.31/MWh × 950 MWh = $16 444. The ISO auction pays the base-load generator $15.00/MWh × 1018 MWh = $15 275. The difference between what the ISO collects from demand and what it owes to generators is a transmission usage surplus, which in this case is $1169. In this scenario, because there is no congestion, the surplus is due to charging demand for marginal losses. Since the ISO is revenue-neutral, it must refund this surplus in some fashion, as discussed in Section 5.7. In the intermediate demand scenario depicted in Fig. 5.2b, while the demand of 1300 MWh is lower than the base-load generator’s capacity, the transmission capacity limit of 350 MW on line 1 ↔ 2 now creates congestion that prevents the base-load generator from meeting the demand on its own.16 This requires dispatching generator B at bus 2 to

EBL

16

The base-load generator congests line 1 → 2 when it injects around 1038 MWh.

Competitive Electricity Markets

202

Bus 1 A

$15

Bus 3

D C

$17.31

950 MW

1018 MW

$16.13

B

Bus 2 Fig. 5.2a. Off-peak scenario.

Bus 1 A

$15

Bus 3

D C

$20.36

1300 MW

1227 MW

$20

EBL

B Bus 2

190 MW

Fig. 5.2b. Intermediate scenario.

Bus 1 A

$15

Bus 3 $41

D

E

C

197 MW 1600 MW

1288 MW

$54.49

B 250 MW

Bus 2 Fig. 5.2c. Peak scenario.

provide “counterﬂow” on line 1 ↔ 2 that can resolve the congestion and also meet the demand at bus 3. In the day-ahead auction, as noted above, this congestion management happens simultaneously in the solution of the auction. The intermediate generator has a start-up cost of $1000 that is considered in the auction’s decision to commit this generator. However, as noted above, this cost does not enter into the calculation of the LMPs. The prices at buses 1, 2, and 3 are $15.00/MWh, $20.00/MWh, and $20.36/MWh, respectively. Using bus 1 as the slack bus for the purpose of calculating LMP components, the energy

Design of US Wholesale Energy and Ancillary Service Auction Markets

203

component of the price is $15.00/MWh, while the loss (congestion) components are $0 ($0), $1.18 ($4.82), and $3.19 ($1.17) per MWh, respectively at the three buses. At these intermediate scenario prices, the ISO auction charges the demand at bus 3 $2036/MWh × 1300 MWh = $26 468. The ISO auction pays the base-load generator $1500/MWh × 1227 MWh = $18 411, and the intermediate generator $2000/MWh × 189 MWh = $3793. The surplus collected by the ISO is now $4264, which is due in this scenario both to marginal loss charges and to marginal congestion charges. Using the above LMP components, the congestion surplus portion of this total surplus is $2163. This means that the loss component is $2101, equaling the sum of the loss LMP components times the net withdrawals ($3857) minus the energy LMP component times the net losses ($1756). However, this division is arbitrary. If the LMP components were instead based on using the load bus as the swing bus, the estimates of the surpluses results would have been different. These calculations are shown in Section 5.7.4.17 In the peak demand scenario in Fig. 5.2c, both the base-load generator and the intermediate generator are operating at the highest output possible given the congestion on the system, but meeting the demand of 1600 MWh requires using one or more peaking units, which in this example are located close to the load at bus 3. Because generator A is still not operating at full output, it remains a marginal generator and the price at its bus remains at $15.00. However, generator B is operating at full output and the price at its bus has risen above its offer price, to $54.49/MWh. This price is the value (in terms of reduced operating cost of the other operating generators) of a hypothetical additional MW (or increment) of power injected at bus 2. The ISO auction has the choice of three supply offers at bus 3, each with different energy and start-up prices. Each of these offers alone is sufﬁcient to meet the demand. The auction unit commitment decision is to pick the generator offer that either (a) minimizes the total offer cost of both energy and start-up if demand is not price responsive or (b) maximizes social welfare deﬁned as the difference between buyer surplus and seller surplus (see Appendix 5A). Mathematically, this is called the auction “objective value.” To illustrate the commitment decision, the results of each choice, including the resulting LMPs, are summarized in Table 5.6. As can be seen, the minimal objective function value is to dispatch the virtual supply offer. This offer has a higher price than generator C, but has no start-up cost. The start-up cost of generator C is high enough that its lower energy offer is not sufﬁcient to yield a lower total cost of energy and start-up. Generator D has a lower start-up cost than generator C, but its higher energy cost results in a higher total supply cost. Note that total payment to sellers is higher than their total offer costs because the LMPs can be higher than the generator offer price at a location. Put another way, some generators are “inframarginal”: in this example, generator B at bus 2. The virtual supply offer is priced in this example to displace Gen C. The expectation of the virtual supplier is that in real-time, Gen C will set the price, allowing the virtual to sell back its position at a lower price than what it has been paid in the day-ahead market. The example is continued in Section 5.5 to show the conditions under which the virtual supply offer makes positive revenues or faces a ﬁnancial loss.

EBL

17 For example, a 1 MW load increment at bus 1 would be met by decreasing load at the “distributed load slack” (just bus 3) by 0.827 MW; thus, losses would be lowered by 0.173 MW, which at $20.36/MWh is worth $3.52.

204

Competitive Electricity Markets

Table 5.6. Comparison of unit commitment choices for peak demand scenario

Energy offer price ($/MWh) Start-up offer price ($) Nodal price at bus 1 ($/MWh) Nodal price at bus 2 ($/MWh) Nodal price at bus 3 ($/MWh) Total demand payments for Energy + start-up ($) Total supply payments for Energy + start-up ($) Auction objective function value ($)

Commit Gen C

Commit Gen D

Commit Virtual E

40.00 2000 15.00 52.80 40.00 62 000

50.00 100 15.00 69.75 50.00 80 100

41.00 0 15.00 54.50 41.00 65 600

42 395

46 702

41 015

34 196

34 265

32 393

5.3.7. Comparison of PJM and New York ISO market rules The general description of the auction markets given above is largely consistent across the US ISOs. For example, with respect to the day-ahead markets, all the ISOs operate auction markets for energy with security-constrained unit commitment which calculate hourly LMPs. However, there are many minor and several quite signiﬁcant differences among the ISO market rules and procedures that affect market behavior and market prices. Beginning in this section and continuing in several subsequent sections, some of these differences between PJM and New York ISO will be examined, so as to further illustrate design choices and trade-offs. A thorough examination of the differences between these two markets, many of which stem from the preferences and expectations of the market designers and market participants in each region, is beyond the scope of this chapter. Readers are encouraged to turn to the ISO tariffs, technical manuals, and annual state of the market reports for further comparative analysis.18

EBL

5.3.7.1. General scheduling procedures and energy market rules Turning ﬁrst to the day-ahead energy markets, some of the basic differences between the auctions in PJM and New York ISO lie in the rules for specifying and submitting energy offers, the components of which are listed in Table 5.7. As this brief survey will suggest, there are some ways in which the PJM offer rules are more restrictive than those in New York ISO, and other ways in which the opposite is true. One notable difference is in the ability to modify offers from hour to hour. PJM requires that a single energy offer, speciﬁed as a piecewise linear curve with up to 11 prices, be submitted for the full day (i.e., the same supply function for each hour), whereas New York, which speciﬁes a stepwise linear function with up to 10 steps, allows offer prices to change from hour to hour. In addition, PJM restricts changes in start-up offers to once in every 6 months, whereas New York allows hourly changes. Hence, with respect to variability over time, the PJM offer rules are more restrictive than those in New York. For some generators that have the option to sell into neighboring markets, or potentially to make a bilateral sale within the ISO market, the hourly ﬂexibility to change offers may 18

These are available on the ISO websites.

Table 5.7. Offer and bid components in the PJM and New York ISO short-term markets PJM

A. Supply Offers and Demand Bids Generation Start-Up Price Minimum Generation Energy Block and Price Dispatchable Energy

Demand Shutdown Price (ﬁxed cost) Demand Bid Regulation Capacity Availability Regulation Price Spinning Reserve Price 10-Minute Non-Synchronized Reserve 30-Minute Operating Reserve

New York

Parameters

Variability

Parameters

Variability

$/hour MW, $/hour

Six months

$/hour MW, $/hour

hourly hourly

Piecewise linear function: 10 pieces, $/MWh for the two points of piece, slope of line $/period $/MW MW $/MWh $/MWh n/a n/a

Daily

Stepwise linear function: 11 steps, $/MWh, MW/Step

hourly

$/MW MW $/MW $/MW $/MW $/MW

hourly hourly hourly hourly, day-ahead only hourly, day-ahead only hourly, day-ahead only

EBL Daily Daily Daily Daily Daily n/a n/a

(Continued)

Table 5.7. (Continued) PJM Parameters

New York Variability

B. Physical Generation Characteristics Dispatch Status Start-Up Time Minimum Run Time Minimum Down Time Max. Start-Ups per Day Normal Upper Operating Limit Emergency Upper Operating Limit Temperature-Based Operating Limits

Normal Response Rate Regulation Response Rate Regulation Maximum and Minimum

Emergency Response Rate Reactive Power Capability Physical Minimum Generation Limit

hours, min hours, min hours, min MW MW MW plotted against temperature MW/min MW/min MW

Variability subject to PJM market rules

Parameters

Variability

Whether ISO or self-committed hours, min hours, min hours, min 1–9 MW MW

May vary May vary per commitment period, day-ahead or real-time Static May change over day May change over day

Limited to Combustion Turbines

EBL

Daily; can be increased but not decreased after day-ahead market closes

MW/min MW/min MW

May vary May vary

MW/min MW plotted against MVARs MW

May vary Static Static

Note: Additional details are found in the PJM and NYISO tariffs. Static refers to offer components that remain relatively constant over the life of the offer, but can be changed.

Design of US Wholesale Energy and Ancillary Service Auction Markets

207

reﬂect changing opportunity costs from hour to hour. In the real-time market, as discussed in Section 5.5, these differences in offer ﬂexibility may have other implications for pricing. Another reason for the differences in offer price ﬂexibility is that the two markets have different market power mitigation rules. These rules will be discussed in more detail in Section 5.8. In short, in PJM, if a generation unit is not transmission constrained, it can generally submit any price offer up to $1000/MWh without any consequence, while in New York, there are offer price screens that may result in an offer being mitigated to a reference price. Hence, in unconstrained areas, PJM appears to permit greater offer pricing ﬂexibility than New York. On the other hand, for units that are in transmission constrained locations, PJM automatically mitigates offers to a pre-submitted marginal cost offer if they fail a market concentration screen (PJM, 2006b), while New York applies offer thresholds according to a formula that becomes progressively more restrictive with the number of congested hours, as summarized in Table 5.13. One of the high-level differences between the market designs in New York and PJM is that the former sought to establish a price basis for as many market features as possible, while the latter retained several physical scheduling features that were more consistent, at least initially, with prior utility practice. For example, PJM allows participants to submit non-price energy schedules, whether bilateral schedules or self-schedules. That is, a generator can simply schedule itself to run in PJM, regardless of market price, as long as the schedule does not cause a reliability concern. Similarly, PJM allows for physical reservations for import schedules on its boundary. In contrast, New York has required that all schedules, including imports, have price offers. A supplier can attempt to guarantee that a generator runs, or that an import is accepted, by reducing its offer prices: the start-up price offer and no-load price offer can be reduced to $0, while the energy price offer can be reduced to $0/MWh or a negative price. However, if a negative price offer is accepted and sets the price, then the generator will be paying to run (NYISO, 2004). In the case of imports, the price offer requirement was an issue for many years due to scheduling difﬁculties that it created due to software problems; other problems for imports stemmed from lack of co-ordination between the ISO markets (Potomac Economics, 2003, pp. 68–74). The upshot is that scheduling only through ﬁnancial offers and market prices can occasionally create complications for market functioning, while physical scheduling may result in economic inefﬁciency. Finding the appropriate balance requires continuing reﬁnement of market rules, software, and inter-ISO co-ordination.

EBL

5.3.7.2. Markets for regulation and reserves New York ISO began market operations in 1999 with auctions for energy, regulation, and three operating reserves. In contrast, PJM operated only an energy market for several years, while procuring ancillary services on a cost basis. PJM introduced a market for regulation in 2000, followed by a market for spinning reserve, called synchronous reserve, in 2002 (see discussion in PJM, 2003). However, the design of its ancillary service markets has been different than the markets in New York, both in the sequence of the markets and in their pricing rules. First, the PJM markets for regulation and synchronized reserve clear in real-time (although price offers have to be submitted on the prior day) and so will be discussed in Section 5.5.6 in the context of the PJM real-time market. New York ISO has both day-ahead and real-time markets for regulation and operating reserves that are simultaneously co-optimized with the energy markets. The rules for these

208

Competitive Electricity Markets

markets are quite complicated and not all the details are covered here (see in particular NYISO, 2006). In New York, day-ahead offers for regulation and reserves are subject to the same submission deadlines as energy (New York ISO, 2001b). Regulation offers must include a regulation response rate (MW/min), which can be no less than 1 MW/min, and a regulation availability price, in $/MW. The three types of operating reserves are 10-minute spinning reserves, 10-minute non-synchronized (non-spinning) reserves, and 30-minute reserves, which can be provided from spinning or non-synchronized units. As shown in Table 5.7, suppliers of operating reserves can submit an “availability” offer ($/MW) for each hour of the day-ahead market and must also provide an emergency response rate that will be used during reserve pickup events. The remainder of the offer consists of physical parameters. The maximum capability for different reserves is a function primarily of the energy output of the unit in relation to its upper operating limit and the unit’s ramp rate.19 New York currently has three locations (zones) for the operating reserve markets: the western zone, which is deﬁned as west of the Central–East transmission path; the eastern zone, which is east of the Central–East transmission path, excluding Long Island; and Long Island. Each of these locations has a reserve requirement, with higher requirements in the east where most of the load is concentrated (NYISO, 2006).20 The pricing of operating reserves is done by zone: there is a different price for each zone described above. Within each zone, the price of each type of reserve under normal operating conditions is done in a fashion that reﬂects the hierarchical ranking and substitution properties.21 In addition, when operating reserve capacity is less than reliability requirements, prices are set by demand curves described next.

EBL

5.3.7.3. Scarcity pricing When supply is tight, reﬂecting scarcity, the two ISOs have different procedures for scarcity pricing. PJM essentially allows suppliers to voluntarily raise offer prices, up to the $1000/MWh offer cap, so as to raise market prices. Some suppliers submit “hockey-stick” offers in which a small percentage of the generator’s output is priced at a high price in the event that demand is high enough to require all available generation capacity including 19

In New York, the rules are that spinning reserve MW are calculated as the unit’s emergency response rate multiplied by 10 (minutes); 10-minute and 30-minute non-synchronized reserve MW are the unit’s upper operating limit (normal or emergency), and for synchronized 30-minute reserves, the emergency response rate multiplied by 20 (minutes), which gives the available reserves above the unit’s 10-minute capability. 20 Currently, the most severe contingency in New York state is rated at 1200 MW and the total 10-minute reserve requirement for the system is set at this quantity. Half of this total is required to be 10-minute spinning reserve, with half of the total spinning reserve purchased in the eastern location; the remainder can be either spinning or non-synchronized 10-minute reserves and must all be purchased on the eastern location. The 30-minute reserve is set at 150% of the most severe contingency, equal to 1800 MW, of which 1200 MW is purchased in the eastern location and 600 MW in the western location. 21 Each surplus higher quality reserve offer can fulﬁll the quantity requirement of a lower quality reserve. The price of 10-minute spinning reserve is equal to the shadow price on the constraint for that reserve in the auction algorithm plus the shadow price for 10-minute non-synchronized reserve plus the shadow price for 30-minute reserve. Similarly, the shadow price for 10-minute non-synchronized reserve is the shadow price for that reserve plus the shadow price for 30-minute reserve. Finally, the price for 30-minute reserve is the shadow price on its constraint.

Design of US Wholesale Energy and Ancillary Service Auction Markets

209

short-term emergency capacity. There is one situation in which PJM will administratively raise prices day-ahead. In the condition of “maximum emergency generation,” the ISO ﬁnds that day-ahead bid demand is not met by offered generation at maximum output. The ISO then takes a sequence of steps to balance supply and demand, setting the market energy price after each step to the highest offer price of any generator on-line or demand offer accepted. First, it increases the output of scheduled generation to their maximum emergency output limits. Second, it schedules generators that are designated as only available for such emergencies. Third, it drops any remaining price-sensitive demand bids. Finally, it sheds load. In the ﬁnal step, the market price is set at the higher of $1000/MWh or the highest offer price of a generator on-line. In contrast, New York has established demand curves for operating reserves that administratively set energy, regulation, and reserve prices during reserve shortages. Until the reserve shortage condition, generators can, as in PJM, attempt to raise the energy price to reﬂect scarcity by submitting high price offers, including those submitted speciﬁcally for output at what is designated as emergency levels (however, unlike PJM, almost all non-emergency supply offers are subject to automatic screening for market power and possible mitigation, as described in Section 5.8). But once the reserve shortage is reached, energy prices are administratively determined through co-optimization with the reserve markets. Currently, there are nine reserve demand curves, some for particular zones.22 In each case, the ISO seeks a target quantity (MW) level for each type of reserve and the demand curve affects the price when the quantity falls below the target level. The curves are simply single price points or stepwise functions when the quantity of reserves reaches a particular level; i.e., they are not negatively sloped. The highest such price in the New York system is $500/MWh for spinning reserve. The lower quality reserves are then priced at lower levels. The prices can go no higher than addition of those demand curve prices, so that even if a high availability offer is submitted in anticipation of a reserve shortage, it cannot further increase the market price of reserves. As the comparison of PJM and New York shows, there are at least two fundamentally different approaches to scarcity pricing currently implemented. Other ISOs have proposed different methods and/or pricing parameters. Hence, there are likely to be many subsequent developments and reﬁnements in this aspect of market design.

EBL

22

The nine demand curves are as follows. For spinning reserves the price is (1) $500/MW for eligible operating reserves at quantities less than or equal to the target level and $0/MW otherwise; (2) $25/MW for eligible operating reserves at quantities less than or equal to the target level for the Eastern region and $0/MW otherwise; or (3) $25/MW for eligible operating reserves at quantities less than or equal to the target level for the Long Island zone and $0/MW otherwise. For total 10-minute reserves (spinning and non-synchronous) (4) $150/MW for eligible operating reserves at quantities less than or equal to the target level and $0/MW otherwise; (5) $500/MW for eligible operating reserves at quantities less than or equal to the target level for the Eastern region and $0/MW otherwise; or (6) $25/MW for eligible operating reserves at quantities less than or equal to the target level for the Long Island zone and $0/MW otherwise. For total 30-minute operating reserves the price is (7) $200/MW for eligible operating reserves at quantities less than or equal to the target level minus 400 MW, $100/MW for eligible operating reserves at quantities less than or equal to the target level minus 200 MW, but greater than the target level minus 400 MW, $500/MW for eligible operating reserves at quantities less than or equal to the target level, but greater than the target level minus 200 MW, and $0/MW otherwise; (8) $25/MW for eligible operating reserves at quantities less than or equal to the target level for the Eastern region and $0/MW otherwise; or (9) $300/MW for eligible operating reserves at quantities less than or equal to the target level for the Long Island zone and $0/MW otherwise.

210

Competitive Electricity Markets

5.4. The Reliability Unit Commitment While the day-ahead market with security-constrained unit commitment is now generally accepted as a design feature that improves the interface between forward markets and reliable system operations in real-time, it still presents the ISO with a few problems in this regard. These problems stem essentially from uncertainty over the physical sufﬁciency of the supply and demand cleared in the day-ahead market. First, as noted, when virtual supply offers are introduced into the day-ahead market, they can displace physical supply offers, potentially leaving the ISO uncertain as to the location and available capacity of generators that are preparing to operate for the next day. Second, if there is no requirement that demand bid in day-ahead, and, further, the market allows virtual demand to bid, the ISO may be uncertain as to whether the demand cleared in the day-ahead market is overor under-scheduled with respect to the operating day. Most US ISOs have addressed these potential problems by adding an intermediate stage to the sequence of the day-ahead and real-time markets, sometimes called a reliability unit commitment. In the reliability unit commitment, the ISO takes two primary actions. First, the ISO removes accepted virtual supply offers from the day-ahead generation offer stack to determine its prospective post-day-ahead physical supply (although the physical generators do not have to operate in real-time, they at least have a ﬁnancial inducement to do so). Second, the ISO compares its own next-day load forecast with the demand cleared in the day-ahead market. If the former exceeds the latter, the ISO seeks to ensure that sufﬁcient generation has been started up to meet load. The ISO then commits additional generators to meet its load forecast. In this fashion, the ISO enters the operating day with greater conﬁdence that it has sufﬁcient generation to meet demand. Like some other design features of ISO markets, the reliability unit commitment is best thought of as a market-priced reliability procedure that could eventually be removed in the event that the demand-side of the markets become sufﬁciently price-responsive. In most ISO markets, it is undertaken after the day-ahead market closes but before the realtime market opens. This means that it is usually conducted in the late afternoon of the prior day, with results made available to market participants over the next few hours. As such, it can also be considered the beginning of the real-time market since, as will be discussed below, any generation committed through the reliability unit commitment will be compensated for any energy produced at real-time prices.

EBL

5.4.1. General procedures and pricing rules As reliability unit commitment designs were introduced into the ISO markets, there was debate over the correct offer and pricing rules. The design decisions were largely focused on two questions: •

Should the reliability unit commitment act as a type of forward market, clearing a market for starting generators as well as energy, or should it simply start generation with sufﬁcient capacity to meet forecast demand while letting the real-time market determine whether to produce energy from those units? • If the reliability unit commitment only starts generators, but does not buy their energy, what is the appropriate pricing for such units? The design direction taken by the ISOs on the ﬁrst question was not to compute a postday-ahead forward market for energy, but rather to minimize the cost of start-up offers by generators such that sufﬁcient capacity is postured to meet the residual demand between

Design of US Wholesale Energy and Ancillary Service Auction Markets

211

the day-ahead market and the ISO’s next-day load forecast. In this fashion, the ISOs could minimize the expenditures by buyers associated with the reliability commitment. If the generators are started up, but their energy is not actually used, then the ISO has not over-purchased energy that would need to be bought back at real-time prices. Conversely, if the generators are started up and do produce energy in real-time, then they are paid for their energy through the real-time market. In fact, all ISOs reduce the costs of the reliability commitment further by applying the real-time market revenue sufﬁciency guarantee. That is, generators scheduled in the reliability unit commitment are not paid until the ISO has determined whether they have earned sufﬁcient revenues in the real-time market through energy sales to cover any start-up and no-load costs incurred. The methods for allocating such uplift charges are discussed in Section 5.6. 5.4.2. Numerical example (continued) In this section, the numerical example begun in Section 5.3 is modiﬁed to reﬂect the actions taken by the ISO in the reliability unit commitment. In the example, a new ISO load forecast is substituted for the demand cleared in the day-ahead market auction. These differences in these demands for each time period are shown in Table 5.8. In addition, the ISO removes the accepted virtual supply offer from the peak demand hours. The results are as follows. In the off-peak period, the ISO estimates that demand will be 25 MWh higher than the demand cleared through the day-ahead market. This additional demand is met at lowest cost by the base-load generator that is already scheduled to operate during that hour. Hence, the ISO does not need to make any additional unit commitments for that period. In the intermediate period, the ISO estimates that demand will be 100 MWh higher than the demand cleared day-ahead. Again, this load is met at lowest cost by the two generators already scheduled in the day-ahead market to operate in that period. In neither the off-peak nor the intermediate periods does the ISO need to consider the displacement of physical generation by virtual supply offers. However, in the peak period, the ISO does have to remove an accepted virtual supply offer from the day-ahead market schedule to conduct the reliability commitment. In doing so, the ISO must consider which physical generator offers are available to meet its forecast peak load, which is 10 MWh lower than the peak demand cleared in the day-ahead market, but which is being met day-ahead with the virtual supply. Both peak generators available in this example have sufﬁcient capacity (MW) to meet the ISO’s load forecast, and the ISO only needs the capacity of one of them, so the ISO’s commitment decision is simply to minimize the costs of starting up one of the generators to meet the forecast demand. In this example, it is generator D, which has high energy costs but low start-up costs, that is committed by the ISO in the reliability unit commitment. By scheduling this generator, the ISO is

EBL

Table 5.8. Difference between day-ahead market demand (MWh) and ISO load forecast (MWh)

Day-Ahead Market ISO Load Forecast Difference

Off-peak

Intermediate

Peak

950 975 +25

1300 1400 +100

1600 1590 −10

212

Competitive Electricity Markets

guaranteeing that it will cover the unit’s start-up costs of $100 whether or not it actually produces energy in real-time. This cost now rolls over into the real-time revenue sufﬁciency guarantee, as discussed in sections 5.5 and 5.6. In turn, generator D is now obligated to start up and be ready to produce energy by the peak period of the real-time market. To simplify the example, assume that generator D is a “quick start” unit that does not have a ramp rate that would require multi-period scheduling.

5.4.3. Comparison of PJM and New York ISO market rules All the US ISOs with a day-ahead market conduct a reliability unit commitment and all procure additional capacity by minimizing start-up and no-load costs. The differences in market rules are relatively minor. PJM begins this procedure, which it calls the “secondary resource commitment,” at 18:00 the prior day. This commitment uses the updated offers via the real-time market and any updated information on resource availability to meet the updated PJM load forecast. Following this initial commitment, PJM will do additional resource commitment prior to the start of the real-time market. Any changes to individual generation schedules are provided to generation owners only. New York ISO undertakes its reliability unit commitment over multiple steps. In contrast to PJM (and the other ISOs), New York integrates its initial reliability unit commitment into its day-ahead market unit commitment. This has been described earlier. In the second pass of the ﬁve-pass unit commitment, the ISO clears additional units based on day-ahead offers to meet the ISOs forecast demand. Like the other ISOs, the objective is to minimize start-up and no-load costs. However, because there is no guarantee that sufﬁcient supply will be offered into the energy market day-ahead, following the day-ahead commitment, New York ISO conducts a resource evaluation called “forecast required energy for dispatch.” In this evaluation, it solicits any additional offers that were not submitted to the day-ahead market or represented in day-ahead schedules. At the start of the realtime market, the ISO continues to do supplemental evaluations to ensure that sufﬁcient offers are available for dispatch. Similarly to other reliability unit commitments, units committed are eligible for start-up and no-load payments and for the revenue sufﬁciency guarantee.

EBL

5.5. Real-Time Market The real-time, or dispatch, market encompasses both an auction market for energy and ancillary services and any submitted MW deviations from day-ahead self-schedules and bilateral schedules. The spot auction is a purely “physical” market in that all sales and price-sensitive demand bids cleared through the market embody requirements to produce or curtail consumption. In the real-time auction market, demand that did not clear in the day-ahead market purchases energy and ancillary services from generators that have been started either through the day-ahead market (and have surplus capacity), the reliability unit commitment or through ongoing real-time commitments. If excess demand cleared through the day-ahead market, then the real-time market calculates the prices at which that demand must be “sold back.” Similarly, if supply clears through the day-ahead market, whether backed by a physical generator or a virtual offer, but does not perform in real-time, its output must be “bought back” at the real-time price. In markets with locational marginal pricing, the real-time market also determines marginal congestion and marginal loss charges for any transactions in real-time. These charges are calculated based

Design of US Wholesale Energy and Ancillary Service Auction Markets

213

on the total dispatch, but settled against deviations from the day-ahead market, as shown in Section 5.8.

5.5.1. Market procedures Offers and bids into the real-time market include those submitted into the day-ahead market but not accepted in that market and new ones submitted subsequent to the dayahead market. ISOs differ over when offers and bids are due, as summarized in Table 5.3. The major difference is between markets that collect all real-time offers and bids the prior day (PJM) and those that allow them to be submitted up to some time just prior to the hourly market in question (all other US ISOs). The price components and physical parameters required for offers into the real-time market are generally the same as the offers into the day-ahead market, as summarized in Table 5.7. The real-time market begins market clearing operations at 00:00 of the operating day and closes at 24:00 of the same day. As with the day-ahead market, by convention it is operated as an hourly market in the US ISOs. Hence, there is a market for each hour (e.g., 12:00–13:00) and typically a single hourly integrated price at each location, even if the dispatch price is calculated on a 5–15 minute‘ basis.

5.5.2. Market pricing There are similarities and differences between the day-ahead market and the real-time market with respect to the calculation of LMPs for energy. The similarity is that once calculated, LMPs clear the energy market. However, the procedure for calculating the prices is quite different. Unlike the day-ahead auction market, the real-time market typically has two parallel programs running to schedule and dispatch generators. The ﬁrst begins with the day-ahead and reliability commitments and conducts unit commitment and de-commitment over the operating day, with a look-ahead usually of a few hours. This unit commitment typically only has to consider peaking units that were not scheduled by the day-ahead commitment or whose schedules need to be adjusted. The second is the dispatch program, which takes the commitment decisions as given and adjusts generator output on a 5-minute basis to achieve an optimal dispatch and set LMPs. The better the integration between these two scheduling programs, the more efﬁcient is the dispatch. Also unlike the day-ahead market, in real-time, the ISO both sends its computed dispatch prices and outputs to generators every few minutes (often 5 or 10 minutes) for the subsequent time period and then meters the actual output a few minutes after that. There are thus two possible ways to set real-time LMPs: on the basis of the ex ante dispatch results or by using the metered outputs to calculate prices ex post. Ex ante real-time pricing has the advantage that it is reﬂective of the ISO’s optimal dispatch for the next time period. However, it generally requires penalties for “uninstructed” deviations (i.e., deviations that take place contrary to system operator instructions), because if a generator knows the anticipated price for the next 5–10 minutes, it may choose to over or under-produce in violation of the ISO’s optimal dispatch. Such deviations may cause costs to other parties, by requiring other generators to ramp up or back down (as regulation), or affect reliability. Ex post pricing relies on incentives for generators to follow their dispatch instruction; generally, if a generator over-produces, it lowers the locational price. But because of transmission network effects, some ISOs do not ﬁnd that ex post pricing offers them sufﬁcient

EBL

214

Competitive Electricity Markets

control over the system and rely on ex ante prices along with penalties to maintain an optimal economic dispatch. Another difference with the day-ahead market is that in real-time, while dispatch instructions and LMPs are calculated on a 5–10 minute basis, for convenience ﬁnancial settlement takes place against an average hourly price. Hence, some hourly realtime LMPs do not cover the offer price of some units dispatched during the hour. Any difference, as in the day-ahead market, is made up through the revenue sufﬁciency guarantee.

5.5.3. Markets for regulation and reserves Real-time markets for regulation and reserves cannot be entirely operated in the context of the 5-minute dispatch because they require commitment decisions and re-scheduling based on the dispatch points that generators have moved to over the operating day. Hence, these are operated as hourly markets and typically re-adjusted in each hour based on new offers and the results of the energy dispatch prior hour. The auction market for these products thus runs in parallel to the energy dispatch auction. However, for any energy produced by a generator providing regulation or reserves, the price paid is the real-time energy price.

5.5.4. Scarcity pricing

EBL

Scarcity pricing in the real-time market typically follows similar principles to that in the day-ahead market. Again, because spot demand is largely non-price responsive, and more so in real-time than day-ahead, most ISOs again use a reserve shortage as a proxy for market shortage in real-time (even if no demand curtailments take place). Two signiﬁcant differences between the two markets are in the trigger for scarcity pricing and ﬁnancial settlement. In real time, system operators will undertake many non-market measures to avoid running short of reserves. Hence, there can be some ambiguity about when scarcity pricing is triggered. Another difference is that due to the averaging of real-time hourly prices, high scarcity prices for a few minutes during the hour will be averaged over the hour, resulting in a diluted price signal to the market.

5.5.5. Congestion management Unlike the day-ahead market, real-time congestion management requires physical re-dispatch throughout the day. The system operators use LMPs and dispatch instructions to resolve congestion, but can also take numerous non-market actions. For example, in PJM, prior to generation re-dispatch, the operator takes all available “non-cost” measures to resolve congestion, including PAR adjustments, transformer tap adjustments, MVAR adjustments, switching capacitors/reactors in/out-of-service, switching transmission facilities in/out-of-service, and curtailing transactions that have indicated that they are “not-willing-to-pay” congestion. After these non-cost measures are completed, the system operator sets a “threshold” for each individual congested transmission element, usually 95–100% of the facility rating, which must be respected by the unit dispatch software. This threshold is then re-adjusted by the PJM dispatcher as conditions change.

Design of US Wholesale Energy and Ancillary Service Auction Markets

215

5.5.6. Numerical example (continued) As noted, the real-time market is a residual market, in which deviations from day-ahead market schedules are settled at real-time prices. To maintain the simplicity of the numerical example, only the supply offers submitted into the day-ahead market but not accepted in that market or the generator committed in the reliability unit commitment are considered for additional real-time demand. The prices and quantities in their supply offers are assumed not to have changed. What does change in the example is the real-time, or physical, demand. The results of each demand scenario are shown in the Figs 5.3a–c. As shown in Table 5.9, actual demand is 5 MW lower in the off-peak period than the quantity cleared day-ahead, but is 150 MW and 20 MW higher than the latter for the intermediate and peak periods, respectively. As a result, the physical dispatch and nodal prices are different from the day-ahead market. In the off-peak period shown in Fig. 5.3a, the base-load generator is dispatched at a level 5.76 MWh lower than the quantity settled in the day-ahead market. Since the generator remains marginal, the price at its bus remains the same as the day-ahead market off-peak period. Hence, the generator has to “buy back” 576 MW × $15/MWh = $8645. Buying back this forward position does not cause the generator to lose money, since it did not produce any physical power day-ahead. Similarly, the demand at bus 3 bought excess power day-ahead, which it now “sells back” at a slightly lower price than the day-ahead price at its location, 5 MW × $1729/MWh = $8645. The differences in the nodal price at bus 3 between day-ahead and real-time are due to differences in losses. In this case, the ISO remains revenue neutral. In the intermediate period shown in Fig. 5.3b, generator C, which was not dispatched in the day-ahead market due to the virtual supplier, is committed because demand is higher than anticipated by the day-ahead market. Remember that the reliability unit commitment postured generator D to potentially provide energy in the peak period, and in doing so committed to paying its start-up. However, by the intermediate period of the real-time market, using its look-ahead unit commitment software, the ISO is aware that both generators C and D will be needed for the peak period. Hence, it is economic to use generator C rather than generator D for the intermediate period. In fact, had both generators not been needed for the peak period, it would have been economic to “decommit” generator D following the reliability unit commitment and only use generator C due to its lower energy price. That is, the savings due to lower energy prices would have more than offset the higher cost of starting up generator C.

EBL

Table 5.9. Results of real-time market simulations and deviations from the day-ahead market Scenario

Demand (MWh)

Supply (MWh)

Price at bus ($/MWh)

A

B

1013 1288 1288

250 250

47 200

Deviations from the day-ahead market: Off-peak −5 −6 Intermediate +150 +61 Peak +20 0

+60 0

+47 +200

Off-peak Intermediate Peak

945 1450 1620

C

D

E

17

1 1500 1500 0 1500

0 +17

−197

0 0 0

2

3 1613 5280 6975

1729 4000 5000

−002 +3280 +1526

+1964 +900

Competitive Electricity Markets

216

Bus 1 A

$15

Bus 3 $17.29

D C 945 MW (–5 MW)

1013 MW (–6 MW)

$16.13

B

Bus 2 Fig. 5.3a. Off-peak scenario.

Bus 1 A

Bus 3

$15

$40

D C

47 MW 1450 MW (+150 MW)

1288 MW (+60 MW)

$52.80

EBL

B

250 MW Bus 2 (+60 MW)

Fig. 5.3b. Intermediate scenario.

Bus 1 A

$15

Bus 3 $50

17 MW D C

200 MW 1620 MW (+20 MW)

1288 MW

$69.75

B 250 MW

Bus 2 Fig. 5.3c. Peak scenario.

In the peak period shown in Fig. 5.3c, both peaking generators are dispatched, because peak demand is higher than was cleared in the day-ahead market and also higher than the ISO forecast in the reliability unit commitment. As noted, the ISO committed generator D in the reliability commitment, and in doing so was obligated to pay the generator its start-up costs regardless of whether it provided energy. Now that the generator has been

Design of US Wholesale Energy and Ancillary Service Auction Markets

217

dispatched to provide energy, the ISO must calculate whether its revenues are sufﬁcient to cover its start-up costs. These calculations are shown in Section 5.6.1. In addition, the virtual supplier must “buy back” its position in the period in which it sold virtual energy day-ahead but at the real-time price. In this example, the virtual supplier buys back 197 MW at bus 3 at the real-time price of $50.00/MWh. Hence, the virtual supplier owes back $9850, for a net loss between day-ahead and real-time of (-$50 + $40) × 197 MW = -$1970. If the virtual supplier’s forecast that the real-time price would be $41.00/MWh had been correct, then it would have owed back −$7879, for a net proﬁt between day-ahead and real-time of ($41 - $40) × 197 MW = $197. 5.5.7. Comparison of PJM and New York ISO market rules As with the day-ahead markets, the PJM and New York real-time markets both calculate LMPs for energy and zonal prices for several ancillary services. However, there are again some interesting differences in market design and procedures. 5.5.7.1. General dispatch procedures and energy market rules Beginning with the energy markets, the two ISOs have different offer price rules and trading deadlines. In PJM, the real-time market offer period begins immediately following the close of the day-ahead market at 16:00 and closes at 18:00 the prior day. In addition, like the day-ahead market, only a single supply function offer can be submitted for all 24 hours of the real-time market. In New York, offers and bids are only due 75 minutes before the operating day hour, and the offer price can vary for each hour. Unlike the day-ahead auction, the real-time auction schedule is the result of the interaction of market functions with several constantly updated forecasting and system monitoring systems. These are worth examining in some detail. In PJM, with all offers submitted, the system operators begin calculating ex ante dispatch instructions for the operating hour using its unit dispatch system. Data from three software systems feeds into the unit dispatch analysis: the energy management system calculates the load forecast, area control error, steam deviation, constraint data, unit sensitivities, and state estimator output; another software system updates unit outage data; and market systems software provides generator offers and data on dispatchable transactions (PJM, 2007b, d). In addition, as described below, a parallel program clears markets for regulation and spinning reserves using the unit dispatch data. The PJM energy dispatch auction subsequently clears using the most current data from the various software subsystems. The unit dispatch system calculates a dispatch solution including dispatch rates and generator unit MW every 5 minutes for a look-ahead period. The dispatcher must approve the solution before the data is sent to utilities and generators. The ISO then solves for a state estimator solution every 5 minutes to estimate the actual MW injections and withdrawals at buses. Actual schedules for external transactions are then included and the ISO sets real-time prices on an ex post basis. There are no penalties for deviations. In New York, as with the day-ahead market, in the real-time market, energy, regulation, and operating reserves are co-optimized. The calculation of real-time schedules and prices is implemented through two primary commitment and dispatch programs that exchange data with each other: the real-time unit commitment program, which conducts a 2.5-hour look-ahead with commitment decisions made on a 15-minute basis, and the real-time dispatch program, which establishes a 5-minute dispatch and calculates market prices on a 5-minute basis. The real-time unit commitment begins with the day-ahead

EBL

218

Competitive Electricity Markets

schedule and either commits or decommits units from that schedule based on offers submitted following the day-ahead market, changes in resource availability and updated load forecasts. Its commitment schedule is binding for units that need to be started up or moved to a dispatch point over a 30-minute look-ahead and advisory for units committed beyond that horizon. The commitment decisions are then passed to the real-time dispatch program, which can adjust the output of committed generators to determine an optimal dispatch and calculate prices.23 The dispatch produces ex ante binding prices and quantities every 5 minutes for the next 5- minute period and provides four additional advisory prices and quantities spaced over 5–15 minutes up to 50–60 minutes ahead, depending on the initial period. 5.5.7.2. Markets for regulation and reserves To establish the real-time dispatch, the ISOs have different methods for co-optimizing the offers for regulation and operating reserves with those for energy. As noted, the New York ISO operates a two-settlement system for regulation and operating reserve (NYISO, 2006). The basic rules for these markets were described above in the section on the day-ahead market. New York buyers purchase most of their regulation and reserve requirements through the day-ahead market. The ISO commits additional resources to provide these services in real-time if insufﬁcient MW were cleared day-ahead, units that were scheduled day-ahead are not available in real-time, or additional MW are needed over the day-ahead forecast (only for regulation, which is a function of actual load). The offer rules for the New York real-time markets for operating reserves are largely the same as those for the day-ahead market, with the exception that all availability offers into the real-time market are assigned a price of $0/MW, meaning that their payments will be based on opportunity costs relative to their energy offer. In contrast to New York, PJM’s ancillary service markets clear after the day-ahead market, and before the real-time market (PJM, 2007b). PJM operates hourly real-time markets for two ancillary services, regulation and spinning reserve (called synchronous reserve).24 Suppliers include eligible generators and demand resources. Demand resources are currently limited to providing 25% of the regulation requirement. There are two types of spinning reserve resources: Tier 1 and Tier 2. Tier 1 is any incremental spinning reserve that is already available through the energy dispatch; i.e., from a generator that is already operating and has additional capacity available through ramping. Tier 2 is spinning reserve from generators that are synchronized to the grid but which need to be dispatched to a different operating point than they would be through the energy dispatch (including generators started-up to produce reserves). Regulation is procured in two separate zones in PJM,

EBL

23

The dispatch program’s procedure for normal (i.e., non-emergency) time periods is a three-pass approach. The ﬁrst pass determines an initial set of binding physical schedules for generators that result from co-optimized minimization of the cost of energy, regulation, and operating reserves. This pass assumes that all ﬁxed block units that have been committed by the real-time commitment are at their upper operating limits, whereas all other dispatchable capacity that has not been committed is ﬂexible. The second pass then relaxes the constraint on loading ﬁxed block units and allows them to be ﬂexibly loaded. This pass determines whether the least-cost solution can be found through inﬂexible or ﬂexible loading of such units. The third pass then calculates LMPs with the optimal mix of inﬂexible or ﬂexible block loaded units. However, the third pass does not change the physical schedule determined in the ﬁrst pass. 24 The regulation market began on June 1, 2000; the spinning reserve market began on December 1, 2002.

Design of US Wholesale Energy and Ancillary Service Auction Markets

219

while spinning reserve is procured in four zones, with zonal prices if transmission constraints separate the zones and a single price when they do not.25 Each load-serving entity must buy a pro-rata share of these services. As with energy, buyers can self-schedule using their own generators, contract with a third party, or purchase through the PJM market. Supply offers for regulation and Tier 2 reserve suppliers are due to PJM by 18:00 the prior day and the price cannot vary by hour of the day.26 Also, any units listed as available for these services but without an offer price are entered into the markets as price takers (i.e., with their prices offers set to zero). However, physical parameters can be changed up to 1 or 2 hours before the dispatch hour, as described below. In PJM, the process for scheduling ancillary services and calculating the dispatch similarly begins 1–2 hours prior to the dispatch hour. The ﬁrst step in real-time market pricing is to calculate the prices for regulation and Tier 2 reserves, which are actually done prior to the real-time energy market and are thus ex ante rather than ex post. Since most units will provide both energy and ancillary services, PJM’s market deadlines for submitting physical parameters for both regulation and spinning reserve are within 1–2 hours of the operating hour. The data on ancillary services (including price offers) is then evaluated by PJM using its unit dispatch system software. For regulation, the ﬁnal regulation capability (MW) above and below the regulation midpoint and the regulation maximum and minimum values (MW) must be ﬁnalized 1 hour prior to the operating hour. For spinning reserves, ramp rates and maximum reserve MW are due 2 hours prior to the dispatch hour. This information is then used to estimate the Tier 1 reserve schedules, which are posted 90 minutes prior to the dispatch hour. Reserve availability and offer quantities for Tier 2 resources are due by 1 hour prior to the dispatch hour. Self-schedules are also due by 1 hour prior, with exceptions for units substituted for others that have become unavailable and for units that have only become available during the dispatch hour. The pricing of ancillary services is then conducted through a co-optimization of forecast energy prices for the hour with the offers and parameters submitted for regulation and spinning reserves. The forecast LMPs are the result of a 1-hour look-ahead provided by PJM’s unit dispatch tool. For regulation, PJM calculates a supply stack that reﬂects each regulation units offer and any opportunity costs incurred by not producing energy. The highest merit order unit price becomes the regulation market clearing price for the hour. Similarly to regulation, for the Tier 2 reserves, PJM’s objective is to calculate a supply stack that reﬂects a Tier 2 unit’s offer price for standing by on reserve as well as any opportunity costs that it might incur by not providing energy (a demand resource has an opportunity cost of zero). The formula is as follows:

EBL

Resource merit order price ($/MWh) = Resource synchronized reserve offer + estimated resource opportunity cost per MWh of capability + energy use per MWh of capability.27 25

The PJM regulation requirement is 1% of the PJM peak load for the day. For regulation, this offer restriction was justiﬁed on the basis of market power concerns; see discussion in Section 5.8. 27 PJM applies different formulas for the estimated resource opportunity cost. For condensing combustion turbines, this opportunity cost is calculated as the absolute value of the difference between the unit’s energy offer price and the forecast LMP at the unit’s bus multiplied by the unit’s MW capability with this number then divided by the unit’s synchronized reserve capability. For non-condensing units, this opportunity cost is calculated as the absolute value of the difference between the forecast LMP and the price estimated for the unit’s set point to provide its assigned quantity of synchronized 26

220

Competitive Electricity Markets

The price of Tier 2 spinning reserves in each zone is the highest resource merit price for the operating hour. The prices for regulation and spinning reserves are posted no later than 30 minutes prior to the operating hour. Tier 1 reserves are priced not for capacity on reserve but on the basis of actual output in response to a reserve call. The price is a $50 premium over the LMP for energy. In the event that spinning reserve units are called to provide energy and do not perform, Tier 1 units are credited for the MW that they provide but are not penalized otherwise, while Tier 2 units have to repay for any non-performance by providing the MW shortfall for the next 3 consecutive, same peak days.

5.6. The Revenue Sufﬁciency Guarantee As described above, participants in the ISO markets are subject both to market incentives and to operational instructions that may be required to maintain system reliability. To establish the appropriate incentives for market sellers and buyers to follow both auction schedules and operational instructions, there are two types of rules. First, there is a “revenue sufﬁciency guarantee” that any multi-part offer accepted by the auction will be fully compensated through additional payments if market revenues are not sufﬁcient to meet its offer requirements. These additional payments are then billed to buyers as an “uplift” – an ex post charge assigned on some averaged basis. This rule will be the subject of this section. There is a second type of rule, related to the ﬁrst but less frequently needed, which is that any participant that has sold or bought through the markets and is then requested by the ISO to change its schedule or physical position for reliability reasons will not lose money in doing so.28 In the early phases of market design, the question of how to calculate and allocate revenue sufﬁciency guarantee charges was a quite heavily contested issue.29 The simplest proposition was that such uplift should be paid on a load-weighted share basis by all demand. However, for merchant suppliers that were interested in primarily transacting through the forward contract markets, there was a concern that allocating the uplift to buyers through the spot markets would provide a ﬁnancial advantage to sellers into the spot markets. That is, a generator seeking to create a short-term bilateral contract would have to incorporate its start-up costs into the forward contract price, whereas spot sellers could have their start-up costs spread out over the buyers in the ISO market. Further, bilateral buyers might thus double-pay if they both paid for start-up via a forward contract and for other spot sellers as part of an uplift charge. Some sellers thus asked that their contracted purchasers be insulated from the market-wide uplift charges. The position taken by the ISOs and supported by FERC (e.g., FERC, 1999) was that generators would be started up to provide both energy and ancillary services, such as regulation and operating reserves; since the latter were reliability services that buyers could not easily dis-aggregate from energy, there was a rationale for allocating the uplift to all buyers on a total demand basis.

EBL

reserve multiplied by the quantity of synchronized reserve provided. Finally, the energy use component of the price is calculated as the forecast LMP multiplied by the MW of energy use divided by the synchronized reserve capability. 28 For example, a generator may be asked to ramp down even though the prior real-time LMP at its bus indicated that it should increase production. These types of problem are often related to difﬁculties in integrating the market price calculation with the short-term operational decisions. 29 See, e.g., discussion of New York ISO’s rules in FERC (1999), pp. 42–3.

Design of US Wholesale Energy and Ancillary Service Auction Markets

221

Another design issue concerns the assignment of such uplift to virtual transactions. As noted above, a virtual supplier “sells” energy in the day-ahead market and then “buys it back” in the real-time market. The virtual seller may displace some physical generation offers in the day-ahead market that are then re-scheduled through the reliability unit commitment. In theory, the start-up and stand-by of the physical units may embody some costs that are then passed to the revenue sufﬁciency guarantee uplift, either day-ahead or in real-time.30 In some cases, such virtual positions may cause uplift to be shifted from one party to another.31 Whether such ex post uplift charges should be allocated to the virtual suppliers is a contested design issues in the US ISO markets (see, e.g., Hogan, 2006). There is no simple and accurate way to calculate the actual price impact of virtual sellers on real-time market uplift charges. The most accurate (in a static sense) would be to re-solve the day-ahead and real-time market sequence with and without virtual transactions represented as a means to quantify the difference. This might be a relatively simple analysis for the day-ahead market, requiring one pass of the unit commitment algorithm, but obviously in real-time, it would require re-solving the dispatch algorithm for every 5-minute market clearing to determine which generators were dispatched to a different level due to the removal of the virtual suppliers. The simplest method is to treat all virtual supply MW as effectively demand in the real-time market and to assign uplift charges to such virtual suppliers proportionally to their share of total MW. This approach does not distinguish the impact of virtual suppliers on the uplift. For example, some virtual supply might be accepted day-ahead but not result in any physical generation being scheduled that would not have been otherwise, but would nevertheless face uplift charges on per-MW basis. In the face of such alternatives, another option is to not charge virtuals any uplift on the basis that the economic beneﬁts that result from market price convergence outweigh any cost shifts that they might cause (e.g., as recommended by Hogan, 2006). FERC has generally supported the principle of “cost-causation” – that entities that create costs should be billed for them – and the simple approach of assigning virtual supply uplift charges on a per-MW basis, although it has not precluded the more complicated analysis of calculating virtual supply’s actual impact on uplift and assigning costs appropriately.

EBL

5.6.1. Numerical example (continued) This section applies the principles of revenue sufﬁciency to the numerical example. In Tables 5.10 and 5.11, the total revenues that each generator earns in the day-ahead and real-time markets are compared to its offer price requirements for energy and start-up 30

Similarly a virtual demand bid that clears the day-ahead market may cause a physical generator to be scheduled that would not have been otherwise. In the reliability unit commitment, the ISO uses its own load forecast so that virtual demand can be removed from the schedule. 31 For example, if there is no charge to virtual suppliers, a vertically integrated utility that is an actual consumer of power, i.e., will have metered energy withdrawal in real time, can also shift uplift charges from day-ahead to real-time, if that is ﬁnancially advantageous, by taking a virtual seller position. To do so, it substitutes virtual supply for its actual generation day-ahead, thus reducing any revenue sufﬁciency uplift paid by its day-ahead load. Its physical generators are then scheduled through the reliability unit commitment or in real time, creating revenue sufﬁciency uplift. However, since its day-ahead load has not deviated, all the uplift is billed to net real-time load.

Table 5.10. Day-ahead market calculation of revenue sufﬁciency Revenues ($) Off-peak Generator A 15 275 Generator B Virtual E

Offer Requirements ($) Intermediate

Peak

Total

18 411 3 793

19 316 13 623 8 076

53 002 15 275 17 416 8 076

EBL

Off-peak

Revenue Sufﬁciency

Intermediate

Peak

Start-up

Total

18 411 3 793

19 316 5 000 8 076

– 1 000 0

53 002 9 793 8 076

0 7 623 0

Table 5.11. Real-time market calculation of revenue sufﬁciency (including reliability unit commitment) Revenues ($) Off-peak Generator Generator Generator Generator

A B C D

−87

Offer Requirements ($) (including reliability unit commitment) Intermediate 905 3186 1879

Peak − − 10 000 849

Total 819 3186 11 879 849

EBL

Off-peak

Intermediate

Peak

Start-up

Total

−87

905 1207 1879

− − 8000 849

− − 2000 100

819 − 11 879 949

Revenue Sufﬁciency

0 1979 0 −100

224

Competitive Electricity Markets

shown in Table 5.4.32 The total revenues for each generator in each energy market are the generator’s output (MWh) in each period multiplied by the LMP ($/MWh) at its location, and summed over all periods. This revenue is shown in column 5 of Tables 5.10 and 5.11 (the source data is shown in Figs 5.2a–c and 5.3a–c, and Tables 5.5 and 5.9). The total offer requirement for each generator is equal to the generator’s start-up offer price and the sum of the generator’s output in each period multiplied by its offer price ($/MWh). This requirement is shown in column 10 of Tables 5.10 and 5.11. All generator offers and the virtual offer are included in Tables 5.10 and 5.11, although not each of these offers is eligible for the revenue sufﬁciency guarantee in the example. For example, the base-load generator A is not assumed to have submitted start-up costs to the ISO because it is not starting-up in the period of the daily auction. Similarly, the virtual offer E is not eligible because it has no start-up offer (and is not allowed to submit one). The ﬁnal column in Tables 5.10 and 5.11 shows whether each offer is revenue sufﬁcient. In the day-ahead market, shown in Table 5.10, each supply offer is revenue sufﬁcient, and generator B makes revenues that exceed its offer requirements. However, in the real-time market, shown in Table 5.11, generator D does not earn sufﬁcient revenues in the energy market to cover its offer prices for energy and start-up. It is thus owed $100 by the ISO. As noted, because the ISO is revenue neutral, this uplift payment is charged to some set of market participants. For example, it could be charged to real-time demand. Table 5.9 shows that in the intermediate and peak periods, 170 MWh of demand was present in real-time that was not cleared day-ahead (ignoring the −5 MWh deviation in the off-peak period). Hence, spreading the $100 over this demand on an averaged basis would result in an approximately $0.59/MWh additional charge to each 1 MWh purchase of energy in real-time. If the −197 MWh deviation caused by the virtual supplier was added to the real-time demand, as it is in some ISOs, then the per MWh uplift would be approximately $0.27/MWh.

EBL

5.6.2. Comparison of PJM and New York ISO market rules The ISOs allocate revenue sufﬁciency guarantee uplift in roughly the same fashion, but with some differences. PJM collects the uplift for the revenue sufﬁciency guarantee for energy as a part of its charges for providing operating reserves, presumably because startup and no-load payments are required for units that are turned on to provide reserves (PJM, 2007b; see also discussion in PJM, 2004). PJM calls this payment “operating reserves charges” and applies them in both the day-ahead and real-time markets. PJM allocates such charges day-ahead to the day-ahead demand, including accepted virtual bids (called decrement bids in PJM), and exports; in real-time, any additional charges are allocated to deviations from day-ahead schedules, including virtual supply (called increment bids in PJM) and demand. Similarly to PJM, New York ISO collects this uplift on a pro-rata basis from all wholesale buyers, but through a per unit charge for transmission scheduling. New York ISO also charges virtual supply for any incremental costs that such transactions cause through the real-time revenue sufﬁciency guarantee (NYISO, 2005). 5.7. Pricing and Settlement of Marginal Congestion and Losses Heretofore, locational marginal pricing of energy has been discussed at a general level, noting that such prices in the ISOs typically reﬂect the effect of both marginal congestion 32

In the ISO markets, the revenue sufﬁciency guarantee extends also to the “no-load” component of the energy offer.

Design of US Wholesale Energy and Ancillary Service Auction Markets

225

and marginal losses. In this section, more detail is provided on how the ISOs calculate these components of LMPs and how they collect and dispose of the marginal congestion and marginal loss surplus payments. The computational issue is worth discussing because it relies on certain technical assumptions and the results are sometimes misunderstood. When the transmission constraints that are examined here – capacity limits on transmission facilities and loss factors – affect LMPs, the ISO almost always collects surplus payments through the energy auction. Because the ISO is a revenue neutral organization, it must dispose of that surplus, i.e., refund it to market participants in some fashion. The rules that are devised for such refunds are somewhat different in each ISO. 5.7.1. Computation of marginal congestion and marginal losses As mentioned in Section 5.3.2, an LMP can be disaggregated into three components, using methods discussed below (Schweppe et al., 1988): 1. The price of energy at a slack or reference bus located on the ISO network; 2. The marginal congestion cost associated with delivering energy from the reference bus to another node and; 3. The marginal loss cost associated with delivering energy from the reference bus to another node. The LMP is the sum of these three factors. The difference in the congestion or loss component between two nodes on the system, A and B, equals the marginal congestion or loss charge associated with injecting at A and withdrawing at B (i.e., a bilateral schedule). There are several ways to calculate the LMP components. The following method is the most common. First, the energy component is deﬁned as the LMP at some (arbitrary) “slack” or “reference” bus, or as a weighted sum of LMPs over a set of “distributed” slack buses assuming a ﬁxed set of proportions. The loss component of a bus LMP is the cost of the marginal losses resulting from increasing the load at that bus by 1 MW, assuming that the entire load increase, including incremental losses, are met from the slack bus (or distributed slacks according to the assumed proportions). The loss component may be positive or negative, depending on whether losses or increase or decrease as a result of the load increase. Finally, the congestion component is the difference between the bus’s LMP and the sum of the energy and loss components. Note that these components depend on the arbitrary choice of a slack bus; what is not arbitrary is their sum – the LMP – at each location. In theory, congestion-only payments to ﬁnancial transmission rights depend on this arbitrary choice of slack. However, ISOs have generally gotten agreement from stakeholders on the deﬁnition of the slack buses; for instance, one approach is to use a distributed slack based on the average load distribution (e.g., California ISO, 2005). This is equivalent to assuming that an addition of 1 MW of load at any bus is met by decreasing the original loads at all buses by the same percentage. An advantage of this deﬁnition is that the load-weighted value of the loss component is then, in theory, equal to zero. The deﬁnition of the LMP components allow for a decomposition of the total surplus earned by the ISO into its loss and congestion components. Multiplying the net withdrawal (load minus generation) at each bus by the congestion component and then summing over all buses yields the congestion surplus; then subtracting this from the total surplus gives an estimate of the loss surplus. Note that the loss surplus will then have two components: the sum of the net withdrawals times the loss component minus the total system losses times the energy component.

EBL

226

Competitive Electricity Markets

An alternative approach for separating the two surpluses is to deﬁne the congestion surplus as the sum of the ﬂowgate prices [dual variables for transmission component capacity constraints in the dispatch solution; see Eqs (2) and (3) of Appendix 5A] times the ﬂow, and then the loss surplus is the congestion surplus subtracted from the total surplus. However, the total loss surplus deﬁned in this manner cannot be disaggregated into a bus-by-bus loss component, and so this deﬁnition has not been adopted by any ISO.

5.7.2. Disposal of marginal congestion charge surplus By deﬁnition, the ISO collects congestion surplus payments from buyers whenever transmission capacity constraints bind in the auction. This is because transmission congestion prevents the cheapest generators from operating at full capacity prior to the dispatch of more expensive units. The surplus occurs when the ISO collects more from buyers than it owes to sellers due to such congestion. Although the total dollar amount that the ISO collects in surplus congestion charges varies over time and among the ISOs, it is generally measured as being under 10% of total market participant expenditures.33 The primary mechanism for refunding these congestion surpluses to ISO market participants is through the assignment of ﬁnancial transmission (property) rights, subject to the requirement of simultaneous feasibility (which ensures that the ISO will be revenue adequate).34 In general, ISOs either allocate transmission rights directly to certain market participants (in the United States, the allocation is to “load-serving entities,” which is typically deﬁned as the party that has the contract to serve retail demand) or conduct auctions for the rights and then provide the auction revenues to the eligible market participants (again, in the United States, the load-serving entities). However they are obtained, these transmission rights collect most of the congestion surpluses that the ISO collects on an hourly basis in the day-ahead market (see, e.g., PJM, 2007a, p. 266). Any surplus congestion rents remaining after transmission rights are settled represents uses of the grid by parties that do not hold transmission rights but pay congestion charges. If there is any congestion surplus left over at the end of the year (or the period deﬁned in the market rules), each ISO has rules deﬁning how it is disposed. Some of these rules are discussed below.

EBL

5.7.3. Disposal of marginal loss charge surplus The calculation of marginal losses provides the market with a more efﬁcient dispatch than would be the case if locational prices only reﬂected congestion. However, because losses are usually represented as a quadratic function of line loadings (e.g., Schweppe et al., 1988), the marginal loss charge between two locations will be greater than the average loss charge. The quadratic function implies that the average cost of losses is 33

For example, PJM reports that total congestion costs have ranged between 7 and 10% of total billings between 2002 and 2006 (PJM, 2007a). See also discussion on congestion metrics in Chapter 4. 34 That is, the ISO will collect sufﬁcient congestion charges in each market in which ﬁnancial transmission rights are settled to pay all the outstanding rights. This rule holds as long as the system topology used in the simultaneous feasibility of the rights remains the same in actual market (see Harvey et al., 1997).

Design of US Wholesale Energy and Ancillary Service Auction Markets

227

roughly one-half the marginal cost. Hence, the ISO will always collect surplus marginal loss payments. The disposal of the marginal loss surplus has been a controversial issue in market design, more for reasons of equity than efﬁciency. In general, any method for disposal of this surplus will support efﬁcient scheduling by a particular market participant as long as the method leaves the participant indifferent between accepting the ISO schedule or dispatch and undertaking an alternative, inefﬁcient schedule or dispatch to obtain loss refunds (e.g., by changing its supply offers or self-scheduling).35 However, equity would suggest that the refund method is not entirely arbitrary, since it could involve unfair transfers among market participants based on pre-existing historical contracts.36 Market participants concerned about the uncertainty of the relationship between marginal loss charges and their share of their refund often have sought a tradable loss hedging right similar in principle to ﬁnancial transmission rights for congestion. Because of the non-linearity of line losses, designing such a tradable loss right has not been straightforward. Losses are an example of diseconomies of scale or super-additive costs, such that Ky1 + Ky2 ≥ Ky1 + y2 . Therefore, there is no simple way to decentralize trading of losses, since loss is a function of power ﬂows (unlike transmission capacity, which is assumed to be independent of power ﬂows in the transmission rights model discussed earlier). However, there is less uncertainty about losses than about congestion charges. Therefore, even though average losses through the year can be approximately as costly as transmission congestion, the risks to transactions are less. Therefore, ﬁnancial transmission rights that cover just congestion costs are likely to cover most of the risks that market participants care about. 5.7.4. Numerical example (continued)

EBL

In the examples given in Sections 5.3.6 and 5.5.5, the total LMP surplus collected by the ISO was calculated as the difference between payments by buyers and payments to sellers. In this section, that total surplus is disaggregated into congestion surpluses and loss surpluses for two of the day-ahead examples: the off-peak and intermediate scenarios (in Section 5.3.6). Changes in the total surplus between day-ahead and real-time are shown in Table 5.12. Turning ﬁrst to the day-ahead off-peak case shown in Fig. 5.2a, the prices at buses 1, 2, and 3 are $15.00/MWh, $16.13/MWh, and $17.31/MWh, respectively. If the slack bus is assumed to be bus 1, then the LMP price components (using the CAISO, 2005 methodology) are an energy component of $15/MWh; loss components of $0.00, $1.13, 35

For example, consider a buyer that has the choice between scheduling generator A or B to serve load at C. A has a lower energy offer price (and actual marginal cost) than B but a higher loss charge than B associated with deliveries to C, resulting in a higher LMP at the load bus than if B was scheduled. The efﬁcient schedule is to dispatch B. As long as the loss refund method makes the buyer indifferent between accepting the scheduling of A or B, it will result in the efﬁcient scheduling of B. In this simple case, if the loss refund exactly matched the loss charge for scheduling A, then the buyer would have the incentive to self-schedule A inefﬁciently. 36 In some US regions, notably the Midwest ISO, some generator siting and contractual decisions made prior to the start of the ISO market were not fully reﬂective of actual losses (implying some cost shifting). In such regions, there was the sense that marginal loss refunds should be related to actual losses in some fashion for some period to reﬂect those historical decisions, even if not on a transaction basis.

Competitive Electricity Markets

228

Table 5.12. ISO surplus collection in dollars due to congestion and losses Total Surplus Day-ahead market

Off-peak Intermediate Peak Total

1 169 4 264 24 585 30 018

Real-time market

Off-peak Intermediate Peak Total (net)

1 155 (−14) 23 605 (+19 341) 33 398 (+8 813) +28 140

(change from day-ahead)

and $2.31 per MWh, respectively, at the three buses; and zero congestion components. The loss component at, for instance, bus 2 is calculated by incrementing the load by 1 MW; to meet that load, 1.0873 more MW of generation is needed from the slack. The value of the 0.0757 MW of losses, evaluated at the slack’s LMP of $15/MWh, is $1.13/MWh. As shown in Section 5.3.6, the difference between what the ISO collects from demand and what it owes to generators is an auction surplus, which in this case is $1169 (Table 5.12). In the day-ahead intermediate scenario shown in Fig. 5.2b, there are both loss and congestion surpluses due to the binding transmission constraint. The prices at buses 1, 2, and 3 are $15.00/MWh, $20.00/MWh, and $20.36/MWh, respectively. Using bus 1 as the slack bus for the purpose of calculating LMP components, the energy component of the price is $15.00/MWh, while the loss (congestion) components are $0 ($0), $1.18 ($4.82), and $3.19 ($1.17) per MWh, respectively at the three buses. As discussed in Section 5.3.6, the total surplus collected by the ISO is $4264 (Table 5.12). Using the above LMP components, the congestion surplus portion of this total surplus is $2163. This means that the loss component is $2,101, equaling the sum of the loss LMP components times the net withdrawals ($3857) minus the energy LMP component times the net losses ($1756). As noted above, this division is arbitrary. If the LMP components were instead based on using the load bus (bus 3) as the slack bus (as in the CAISO, 2005 methodology), the estimates of the surpluses results would have been different. Then the energy component would be $20.36 (the bus 3 price), and the loss components would have been −$3.52, −$2.20, and $0.00 per MWh at the three respective buses.37 The resulting congestion components would also be negative, being −$1.84, $1.84, and $0.00 per MWh, respectively. The congestion surplus would then be calculated as (−184×−1227 +184×−190) or $1920. Subtracted from the total surplus of $4264, this yields a loss surplus of $2355. Thus, the loss and congestion surpluses based on bus 3 being the slack are each about 10% different from the values based on a bus 1 slack, above. Finally, if the congestion surplus was instead deﬁned based on ﬂowgate shadow prices, it would instead equal $5.39 (the price for the congested ﬂowgate between buses 1 and 2) times 350 MW (the corresponding ﬂow), or $1886; this would imply a total loss surplus of $4264 − $1886 = $2378. These values are close but not identical to the component surpluses

EBL

37

See also discussion in footnote 17.

Design of US Wholesale Energy and Ancillary Service Auction Markets

229

resulting from using the distributed load slack (bus 3) as the slack bus. As noted, this method is not used by ISOs because it does not disaggregate to the bus level. LMPs are calculated both day-ahead and in real-time, and the value of the ISO surpluses may change between the two markets, as shown in Table 5.12. For example, in the off-peak scenarios shown in Figs 5.2a and 5.3a, since there is no congestion, all the surplus collected by the ISO is due to losses. There is less demand off-peak in real-time than day-ahead, so the ISO effectively “owes back” $14 in loss surplus that was assigned to day-ahead buyers after the resettlement of the sellers’ and buyers’ real-time positions (as discussed in Section 5.5.5). In practice, the ISO surpluses are not disbursed or refunded on an hourly basis, but are aggregated for later disbursal. In the intermediate and peak scenarios, there are also congestion surpluses due to the binding transmission constraint. In the day-ahead intermediate market scenario, the congestion and loss surpluses are fairly close in value. However, in the real-time intermediate market scenario, due to the increase in demand that requires the dispatch of the more expensive generator at bus 3 and continued congestion causing an increase in the price differentials between buses 1 and 3 (but also creating a negative differential between buses 2 and 3), the total surplus greatly increases relative to the day-ahead surplus. The ISO thus collects additional surpluses in real-time, which are recovered from real-time buyers. A similar result is seen in the peak scenario. The interpretation of these results must be done carefully. If market participants hold ﬁnancial transmission rights and are concerned that the settlements of such rights dayahead will not collect sufﬁcient revenues compared to congestion charges in real-time, then they can use virtual bids and offers to effectively shift their congestion hedge to real-time. This was discussed in Section 5.3.1. With regard to changes in marginal loss surpluses between day-ahead and real-time, these surpluses in total are refunded to market participants, regardless of which market they occurred in. Some ISOs have speciﬁc rules to account for differences in magnitudes between the day-ahead and real-time loss surplus so as to ensure that these differences do not create inefﬁcient scheduling incentives when the loss refunds are determined.

EBL

5.7.5. Comparison of PJM and New York ISO market rules As with many areas of auction market design, there are some differences in the rules in PJM and New York regarding dispersal of ISO congestion and loss charge surpluses. Beginning with congestion surpluses, in both markets, these are used initially to pay the set of awarded ﬁnancial transmission rights. Any residual congestion surplus is then subsequently refunded to market participants. As of this writing, in PJM, the existing transmission rights are paid in full only if the ISO collected sufﬁcient congestion revenues to do so; otherwise, payments are pro-rated.38 At the end of each month and over the course of the period covered by annual allocations of ﬁnancial transmission rights, PJM then distributes any residual congestion surplus in ﬁve stages (PJM, 2007b, p. 46). In the ﬁrst stage, any surplus is allocated to holders of ﬁnancial transmission rights that were deﬁcient with respect to their target allocation for the month. In the second stage, any remaining surplus is allocated to ﬁnancial transmission right payment deﬁciencies in prior months of the year. Third, any remaining surplus (after stages one and two) is carried 38

This rule is due to change in 2008, at which time PJM will pay all transmission rights in full regardless of shortfalls in congestion revenues, through an uplift charge.

230

Competitive Electricity Markets

forward to the subsequent month and distributed in the same fashion. Fourth, at the end of the year, any remaining surplus is used to compensate for any deﬁciency in payments to auction revenue rights. Finally, any remaining annual surplus is distributed to entities that had paid PJM transmission access charges that year in proportion to their demand charges or MW reserved capacity for transmission service into, out of, or through the transmission system. New York ISO has a different method for settlement of ﬁnancial transmission rights in the event of a congestion rent shortfall, and hence the rules for disposing of any residual congestion surplus after cashing out the transmission rights are also different. Unlike PJM, in the event of a shortfall in congestion revenues, payments to the ﬁnancial transmission rights are made wholly by transmission owners in New York via a pass-through to their retail rates (sometimes called “full funding” of the rights). Because of this rule, any residual congestion charge surplus is disbursed to transmission owners. With respect to losses, both PJM and New York ISO implement a marginal loss calculation in locational marginal pricing. New York ISO’s surplus, which it calls the “residual loss payment,” along with surpluses collected in other ways (with the exception of congestion cost surplus), is credited against what transmission customers are billed as aggregate ISO costs, which include operational costs and the costs of implementing certain ISO programs, such as demand response payments (NYISO, 2001a, Rate Schedule 1).

5.8. Market Power Monitoring and Mitigation

EBL

The basic deﬁnitions and procedures of market power monitoring and mitigation were introduced in Section 5.2. Here more detail is provided, in particular on the ISO auction rules for economic withholding. ISO markets typically encompass sufﬁcient geographic territory and enough suppliers that generation market concentration is relatively low by most market-wide measures, except in two cases: ﬁrst, when transmission constraints bind causing ISO market concentration to rise in sub-regions, and second, when supply is scarce and the lack of demand elasticity allows even small suppliers to raise prices. Both of these market conditions are typically intermittent, and their frequency is a factor in creating the mitigation rules. For example, if a transmission constraint causes a persistent, concentrated sub-market to emerge, sometimes called a “load pocket,” then a speciﬁc remedy may be required for that sub-market, e.g., requiring some generators to sign cost-based contracts.

5.8.1. Identifying and mitigating exercise of market power In the US ISOs, three actions are typically identiﬁed as the exercise of market power: physical withholding, economic withholding, and uneconomic production. Physical withholding refers to any withdrawal of physical generation capacity from the market, including both units that are self-scheduled and those offered into the auction, that could be construed as an attempt to raise the market price. This could result from unscheduled outages that cannot be justiﬁed ex post, changes in a generator’s physical parameters in its offer that cause it to reduce its output, and failure to respond to an ISO’s dispatch instructions. Forced outages and scheduled maintenance are obviously not considered physical withholding. Physical withholding is typically simple to detect, but can be difﬁcult to prove as intentional exercise of market power in the absence of clear evidence. Hence,

Design of US Wholesale Energy and Ancillary Service Auction Markets

231

ISOs have adopted certain measures, such as a percentage share of a generator’s available capability withheld or percentage deviation from a dispatch instruction, before physical withholding is identiﬁed as a market rule violation. Mitigation measures are ex post, and typically involve either revising the generator’s offer or applying penalties. Economic withholding refers to raising offer prices substantially above marginal cost, including opportunity cost, so as to affect the market clearing price. Typically, an offer that is economically withheld is one that has raised its price sufﬁciently to not get picked by the auction. However, in electricity markets, short-term demand is largely inelastic – i.e., not price sensitive – in which case there will be times when all offers are picked and any price can theoretically clear the market. One situation in which this could be the case when demand is inelastic is if one or more suppliers are “pivotal,” meaning that if they withdraw all their available capability the market could not clear. Most ISOs have implemented a threefold approach to mitigation of economic withholding in the energy market. The ﬁrst component is an absolute offer cap on day-ahead and real-time energy, which currently stands at $400/MWh in California and $1000/MWh in the eastern US ISOs. No offer can exceed this cap, but a locational market price can be higher than the cap due to network effects. A second component, present in some ISO markets but not in others, is offer caps outside the transmission constrained area that are below the absolute cap. In some ISOs, such as New York ISO as described later, these have been implemented in a two-step fashion sometimes called a “conduct-impact test.” In the ﬁrst step (the conduct test), an offer is screened to determined whether it has violated a percentage increase, or absolute dollar amount increase, from a “reference” accepted offer price. The reference offer is typically an average of prior accepted offers and is understood to be a proxy for a competitive market offer. In the second step (the impact test), the ISO determines whether offers that violated the ﬁrst step also increased the LMPs by some additional threshold. An offer that violates both steps is mitigated to its reference offer price and the auction is re-run. The third component is offer caps for persistently transmission constrained locations, where presumably generation market power is more prevalent. These tend to be more restrictive than the offer caps for locations that are less frequently transmission constrained. They are also strict caps rather than the screens discussed above. ISOs have adopted different approaches to monitoring and mitigating the price offers for start-up and no-load. In some ISOs, these are restricted to being modiﬁed infrequently, while in other ISOs, where they can change daily, they can be subject to similar offer caps to energy. There are also offer caps in the ancillary services markets; in some cases, where there is administrative scarcity pricing for operating reserves, those prices effectively set the reserves market price and through simultaneous co-optimization of energy and reserves also set the energy LMPs.

EBL

5.8.2. Contractual remedies There are typically a number of generators in ISO markets that have locational market power because they are needed to operate for reliability purposes or persistently to resolve congestion. Some such generators were built to provide transmission support. Others are older generators that are only intermittently used for peak hours but which cannot recover ﬁxed costs through the market. For such units, which are sometimes called “reliability must run” units, a contractual solution may be required. Such contracts may be cost-based or pay the higher of a market price or a contract price.

232

Competitive Electricity Markets

5.8.3. Comparison of PJM and New York ISO market rules PJM and New York ISO have a number of differences in their market power monitoring and mitigation rules, stemming from their regulatory histories and aspects of their subsequent market development. The most prominent difference is that PJM uses mandatory cost-based offers as a basis for mitigating transmission constrained supply offers (if needed), while New York ISO has a quite different approach under which the system is subdivided into pre-speciﬁed “constrained areas” in which there are formulas for offer caps that depend on the frequency of congestion and areas outside the constrained areas, where a conduct-impact test is applied. These differences will be explained below. PJM began its centralized auction market for energy in April 1998 and operated it for a full year with all supply offers capped at marginal costs.39 When the bid-based energy market was begun in April 1999, PJM continued this method of mitigating supply offers when an energy offer was from a generator’s whose output was altered due to transmission congestion, called an “out-of-merit” generator. Subsequently to beginning its markets, PJM also tightened the rules for start-up and no-load. Currently, a generator can choose between cost-based and price-based offers for these offer components. If it chooses cost-based offers, these can be adjusted daily. However, if it chooses price-based offers, these can only be adjusted twice a year, during enrollment periods (a generator can also switch between cost- and price-based offers in these enrollment periods). Other rules have addressed physical offer parameters, which can provide a generator with market power with additional means to affect market clearing prices. For example, in PJM, the market monitor found in the summer of 1999 that during peak demand hours when the ISO requested all generators to produce on an emergency basis (called “maximum generation emergency alert” in PJM), certain generators anticipated the shortage and increased their minimum run times to the full day. This allowed them to get paid their offer price (including start-up) in hours of the day when the market price had subsided below their offer. The market rule change to address this issue required that a generator’s total offer, including payments for start-up and no-load, could not exceed $1000/MWh during the speciﬁc hours of the emergency. In any other hours that the generator would operate due to its minimum run time, it would be a price-taker in the market and would not be eligible for additional revenue sufﬁciency payments (FERC, 2000). Sometimes, market power rules are overly restrictive. In PJM, for generators subject to offer caps due to being out-of-merit, concern grew that particularly for persistently offer-capped units, energy market prices were not sufﬁcient for such units to recover long-term variable costs. In PJM, this lead to a reduced application of offer caps through two measures. First, PJM developed a new test, called the “three pivotal supplier test,” to determine when a transmission constraint was creating a truly uncompetitive market behind the constraint. The offer cap is not applied to a generator that affects a particular constrained transmission path if there are three or fewer suppliers that are jointly pivotal with respect to the constraint and if the owner of the generator when combined with

EBL

39

This was because FERC rejected PJM’s ﬁrst market proposal in 1997 which did not provide sufﬁcient information about generation market power to satisfy regulatory requirements (FERC, 1997). However, while PJM undertook its market power analysis, FERC allowed it to start market operations in April 1998 with the condition that market power was mitigated in the interim using offers capped at marginal cost. The market with liberalized supply offers began one year later.

Design of US Wholesale Energy and Ancillary Service Auction Markets

233

the two largest other suppliers affecting the constraint is not pivotal (PJM Tariff, Section 5.6.4). Second, PJM allows generators that are frequently mitigated to recover a greater percentage over and above their incremental cost offer cap.40 PJM has also established market power mitigation rules for its regulation and reserves markets. At the start of the regulation market in 2000, PJM expressed some concern about the market concentration (i.e., the low number of potential sellers) and as a result was allowed to establish the requirement that regulation price offers cannot vary by hour and later also that they would be subject to a $100/MWh cap. Moreover, due to market concentration some generators in PJM can only submit cost-based offers for regulation. In the spinning reserve market, several zones have caps based on cost-based rates. New York ISO initially began its markets in November 1999 with generation offer restrictions only within New York City, an obvious load pocket, called a “constrained area.” FERC approved this approach based on New York ISO’s demonstration that the market outside the constrained area was sufﬁciently un-concentrated. However, in 2000, New York ISO and its external market monitor developed the “conduct-impact” approach to screening supply offers described above, which has subsequently been adopted directly or in modiﬁed form by the other ISO markets, with the exception of PJM. As a component of this method, New York developed a market-based approach to measuring a benchmark competitive offer (and hence to avoid requiring generators to submit marginal cost offers, as in PJM). The basic method is to use the average of a generator’s accepted offers in the prior 90 days as a reference price. If there is not a sufﬁcient history of accepted offers, then there are other methods for setting reference prices. The offer caps under this rule are shown in Table 5.13. As in PJM, the rules differ between constrained areas and those in largely unconstrained areas. This is not surprising since a market-based reference offer would reﬂect the market concentration in the constrained area. Hence, as shown in Table 5.13, the New York ISO resorts to offer caps that are a declining function of the average price in the constrained areas and the number of hours that the area is congested. Like PJM, New York ISO has also had some experience with overly restrictive mitigation rules. Generators in New York City was initially subject almost continuously to offer caps set roughly at the level of their variable production costs. In 2004, these rules were revised to conform to the conduct-impact approach used in other parts of the system (Potomac Economics, 2005). This reduced the frequency of mitigation and allowed prices to rise to levels more reﬂective of scarcity within the area.

EBL

5.9. Other Topics in ISO Market Design and Implementation This section reviews several other topics in ISO market design and implementation that have been prominent in the United States in the period under review. The ﬁrst of these is the interaction between the daily energy auction markets and longer-term markets and other functions undertaken by the ISO. The second is the continued existence of market “seams” that include boundaries of market operations that do not conform to natural 40

Speciﬁcally, for units that are offer capped for (a) between 60% and 70% of their run hours, the offer cap is either incremental cost plus 10% or incremental cost plus $20/MWh; (b) between 70% and 80% of their run hours, the offer cap is either incremental cost plus 15% or incremental cost plus $30/MWh, and (c) 80% or more of their run hours, the offer cap is either incremental cost plus 10%, incremental cost plus $40/MWh, or unit-speciﬁc going forward costs in agreement with the ISO.

234

Competitive Electricity Markets

Table 5.13. Thresholds for identifying conduct that may lead to economic withholding in New York ISO Offer Components

Outside Constrained Area

Within Constrained Area (excluding New York City)

Energy and minimum generation offers: Day-ahead market

Lower of 300% or $100/MWh increase; Offers below $25/MWh excluded.

When constraint binds, lower of (a) thresholds for outside Constrained Area, or (b) a threshold calculated as follows: (2% × Average Price × 8760)/Constrained Hours, where Average Price is the average day-ahead market price in the Constrained Area over the prior 12 months (adjusted for fuel price changes) and Constrained Hours is the total number of hours in the day-ahead market over the prior 12 months in which any transmission interface or facility leading into the Constrained Area where the generator is located had a shadow price > 0 in any interval.

Energy and minimum generation offers: real-time market

Lower of 300% or $100/MWh increase; Offers below $25/MWh excluded.

Same as day-ahead calculation, with the Average Price and Constrained Hours calculation being made using real-time market data. The Average Price is also adjusted for out-of-merit generation dispatch as feasible and appropriate.

Start-up price offer

Increase of 200%.

Regulation and operating reserves offers

Lower of 300% or $50/MWh increase; offers below $5/MWh excluded.

Lower of 300% or $50/MWh increase; offers below $5/MWh excluded.

Time-based offer parameters (including start-up times, minimum run times, minimum down times)

Increase of 3 hours or increase of 6 hours for multiple time-based bid parameters.

Increase of 3 hours or increase of 6 hours for multiple time-based bid parameters.

Offer parameters in units other than time or dollars (including ramp rates and maximum stops)

Increase of 100% for parameters that are minimum values; 50% decrease for parameters that are maximum values.

Increase of 100% for parameters that are minimum values; 50% decrease for parameters that are maximum values.

EBL

Increase of 50%.

Source: New York ISO Market Services Tariff, Attachment H. Available at: www.nyiso.com

operational boundaries as well as differences in market design and other factors that may result in economic inefﬁciency. The third is the important role of software design and development in expanding the scope of the auction markets and capturing opportunities for efﬁciency.

Design of US Wholesale Energy and Ancillary Service Auction Markets

235

5.9.1. Longer-term ISO markets and operational/planning functions In addition to the energy and ancillary service markets described in this chapter, which take place on a daily and hourly basis, the ISO also operates markets and undertakes reliability, operational, and planning functions that take place on longer time-frames, from several days to several years. With respect to markets, these include most notably auctions for generation installed capacity and ﬁnancial transmission property rights. Table 5.14 lists several of these market and other functions and their respective time-frames. The design of these longer-term ISO markets is discussed in several chapters in this volume and will not be reviewed here. Notably, none of the ISOs currently operate forward energy auction markets on a longer time-frame than the day-ahead period discussed in this chapter. These longer-term markets and functions have interactions with the daily markets, both intended and unintended. For example, the capacity product is the operable MW of a generator and is sold separately from forward or spot energy. In the ISO markets, there is no need to purchase energy and capacity from the same generator. However, most of the capacity products are designed implicitly or explicitly as options for the ISO to call on energy from a generator in the event of shortages. If a capacity generator is called on for energy by the ISO, then it must curtail any sales outside the ISO market and provide real-time energy to the zone for which it has been designated as a capacity resource.41 As ISOs have found out, it is important to design capacity markets that have sufﬁcient incentives or penalties to enforce this call option. Another type of possible linkage, this time unintended, between forward and spot markets is between holdings of ﬁnancial transmission rights and energy market behavior. A market participant that both owns generation and can collect congestion revenues from its transmission rights may have the

EBL

Table 5.14. Longer-term ISO markets and operational/planning functions Time-frame

Market Functions and Operations

Reliability and Planning Functions

Multi-year

Forward markets for installed capacity Allocation and auction of multi-year ﬁnancial transmission rights

Regional planning and expansion (transmission, generation, and demand response)

Annual, monthly

Allocation and auction of annual, seasonal and monthly ﬁnancial transmission rights

Co-ordination of planned transmission and generation maintenance

Weekly

Scheduling of generation with more than one-day start-up times

Load forecasting

41

Hence, under current designs, a capacity contract is not equivalent to a forward energy contract – i.e., physically linked to a particular buyer of power – because the ISOs are not equipped to implement a priority ranking among wholesale buyers in the event of a load curtailment. Instead, the capacity product is deﬁned zonally, as an option to deliver power from a generator to a zone. For example, PJM conducts a “deliverability test” for generators requesting to be capacity resources using a power ﬂow model that evaluates how the power from that generator will diffuse in a region of the market. When transmission constraints limit the ﬂow of the generator in this test, it must pay for upgrades to fully qualify as a capacity resource. Once it meets the eligibility requirements, a capacity resource is then invested with energy market obligations in exchange for capacity payments, as noted earlier.

236

Competitive Electricity Markets

incentive to alter its output so as to maximize the revenues from both energy and its transmission right revenues. This would result in an inefﬁcient dispatch. Such activity has been observed in the markets, but is generally rare. 5.9.2. Market seams A market “seam” refers to differences between methods of power system operations, market designs, and rules for crossing market boundaries (i.e., as an importer or exporter of power) that create transactions costs or externalities across the boundary. In the United States, the ISO markets can have seams with other ISO markets or with the regions that do not have organized markets. In each case, the primary issue across the seam is transmission scheduling, congestion management, and unscheduled ﬂows (such as “loop ﬂows”). FERC has sought to reduce or eliminate these market seams over the years. In late 1999, it issued a rule that encouraged and provided incentives for all utilities in the United States to join large Regional Transmission Organizations (RTOs) on a voluntary basis (FERC, 1999). A subsequent FERC initiative to merge PJM, New York, and New England on the basis that they already shared a similar market organization failed in this period due to regional differences. In 2002, with RTO formation lagging, FERC proposed a “standard market design” for all utilities in the country that largely followed the basic ISO design described in this chapter (FERC, 2002). However, this proposal met strong political and industry resistance in some regions of the country and was formally revoked in 2005. Hence, for the period covered by this chapter, market seams, both between ISO markets and between the ISO markets and the purely bilateral markets, have remained salient concerns from both an operational and economic perspective. Had the standard market design rule been implemented, many of these seams problems would have been resolved.

EBL

5.9.3. Developments in market software In practice, software has been a limiting factor in the development of efﬁcient market designs. The existing ISO software and data systems are a result of market start-up decisions as well as patches resulting from continual change and improvement. Consequently, changes to a single software system may require changes to many software and data systems. Currently, there is a signiﬁcant backlog of improvements in each ISO. In part the backlog is due to the extensive testing and changes necessary to install new software modules. Hence, there are still signiﬁcant efﬁciencies to be gained from the standardization of data systems so that when improved software is developed in one area it can be easily tested and put into production elsewhere. Nevertheless, in some areas, ISOs and commercial vendors have been able to introduce innovations that demonstrate the value of their investment in software and growing technological expertise. These innovations have produced economic beneﬁts. One example is the search for faster solution times and closer to optimal solutions for the unit commitment auctions that take place day-ahead and then over the operating day. In the past, the Lagrangian relaxation solution algorithm used by utilities and then by the ISOs was known to reach sub-optimal solutions, but the size of the problems that it could solve and its solution times had been reduced over the years. In theory, mixed integer programs could achieve an optimal solution but had difﬁculties with solution speed for large problems. In recent years, these solution time issues have been reduced such that ISOs can

Design of US Wholesale Energy and Ancillary Service Auction Markets

237

now deploy mixed integer programs for the auction markets (Johnson et al., 1997; Hobbs et al., 2001; Guan et al., 2003). In addition, mixed integer programs allows a much more detailed and full speciﬁcation of the unit commitment problem and relieves the market co-ordinator of many of the simplifying assumptions that are necessary using Lagrangian relaxation.42 These advantages and the tractability of mixed integer programming algorithms have led several ISOs to introduce or test mixed integer program-based implementations over Lagrangian relaxation. Due to the computational complexity of unit commitment problems, ISOs which implement mixed integer program based algorithms tend not to solve their unit commitment problems to optimality due to limitations on solution times.43 Streiffert et al. (2005) note that the enhanced modeling capabilities of mixed integer program allows the ISO to deal directly with a number of constraints that were very difﬁcult to model in Lagrangian relaxation. PJM has implemented mixed integer programming for its day-ahead market with estimated savings of $54 million per year due to efﬁciency improvements in the auction solution. While savings from the more precise solutions provided by mixed integer programs may be a small percentage of total generation costs, i.e., 1–4%, a 1% savings in generation costs translates into a $1–$2 billion annual cost saving in the United States alone. 5.10. Extensions of the Market Design The US ISOs have been operating spot energy auctions with locational marginal pricing since 1998, and regulation and operating reserves markets with various designs for almost as long. As discussed in this chapter, market designs and software are continuously being reﬁned in response to various factors, including incomplete markets, the introduction of new technologies and software, and also due to the greater penetration of some existing technologies that create new operational requirements, such as demand response and wind energy. The auction designs described in this chapter should be adaptable, such that the basic organizational and pricing principles will remain appropriate as technological changes take place. Among the near-term design challenges relevant to this chapter is more “complete” pricing of ancillary services, such as reactive power. Currently, many systems are dispatched without using a full AC optimal power ﬂow, and impose overly restrictive voltage levels. Reactive power is not priced or is priced inappropriately. The design question is whether ISOs should sign long-term contracts for reactive power that include obligations

EBL

42

Even if the mixed integer program algorithm times out before ﬁnding an optimum, one is still left with a primal-feasible solution and a bound on the optimality gap. These intermediate solutions are often found within the same amount of time a Lagrangian relaxation-based algorithm takes, and typically have optimality gaps of the same size or smaller than Lagrangian relaxation commitments. Furthermore, a mixed integer program-based solution algorithm allows ISOs to easily introduce new types of unit-operating and system constraints to the formulation of the problem, whereas Lagrangian relaxation-based techniques generally require extensive reprogramming of the feasibility heuristics to ensure that the unit commitment satisﬁes all the necessary conditions. 43 PJM, for instance, allows its mixed integer program optimizer to run within a certain period of time or until the optimality gap is below some maximal threshold and uses whatever intermediate integer-feasible solution the solver has found. If an ISO is left to rely on an intermediate integerfeasible but sub-optimal solution, the same issues of generator payoffs, energy pricing, and inequity of the resulting dispatch arise as with sub-optimal Lagrangian relaxation commitments.

238

Competitive Electricity Markets

to perform or whether spot markets for reactive power can elicit a more efﬁcient outcome and overcome concerns about locational market power (Hogan, 1993; Kahn and Baldick, 1994; FERC, 2005). To implement such markets, optimal power ﬂow software needs to be improved and integrated with unit commitment models. Another line of inquiry concerns the more active participation of “dispatchable” transmission elements in the energy spot markets (O’Neill et al., 2005). That is, with sufﬁcient regulatory oversight, both merchant and regulated transmission owners could offer some or all of their transmission capacity into the ISO markets at a price. Another aspect of this development could be the unit commitment of transmission elements; e.g., the ability to decouple a transmission line if that leads to a reduction in the auction objective function.

5.11. Conclusions The US ISO markets, many encompassing multiple states and reaching a geographic scope that was not anticipated just a few years earlier, are a major achievement of electricity regulatory reform. The auction market designs for energy and ancillary services have developed on the basis of both theoretical principles and practical decisions. In some cases, design mistakes were made, but equally the ISOs and their stakeholders have worked to reﬁne the designs and learn from experience, with FERC oversight. This has led to a high degree of convergence on key elements of the designs, such as day-ahead and real-time markets with locational marginal pricing, the reliability unit commitment, and the co-optimization of regulation and reserves; although as this chapter has shown, there remain many design differences between the ISO markets. A lesson of the ﬁrst decade of the ISO markets is that an efﬁcient spot market for electric power that respects economic principles and reliability requirements ends up being rather complicated. There are many products and their pricing and ﬁnancial settlement rules are often difﬁcult for market participants to understand and analyze. This chapter has sought to provide a step-by-step review of many of these market rules and procedures and to explain why design choices were made. More recently, concerns about the costbeneﬁt ratio of implementing such markets have been raised. Among other things, this has prompted calls for simpliﬁcation of the market designs. Although simpliﬁcation where possible (and increased standardization) of market rules in the United States should be a design objective, observers should also note that the complicated design of the daily auction markets is in part because so much of the power system has been exposed to transparent market pricing and procedures. What in prior years were system operational decisions whose costs were internalized by utilities, often inefﬁciently, are now integrated into daily market auctions and generally priced efﬁciently. Transmission usage is priced on the margin and optimized over large regions, resulting in much more efﬁcient use of transmission capacity. Extensions of these locational pricing principles and the application of unit commitment to transmission elements could yet result in even further gains in utilization of the existing grid. However, as discussed, there are several aspects of market pricing that do continue to rely for practical reasons on types of average pricing. Most notably, most ISOs still charge buyers load-weighted average LMPs on a zonal basis rather than the LMPs at their nodes. The current market designs will continue to evolve with reﬁnement of the market rules, changes in technology, and shifting regulatory requirements. The markets will function better if technology and consumer interest (or regulation) allow a robust demand response to emerge. At least some of the regulatory and reliability aspects of the market designs,

EBL

Design of US Wholesale Energy and Ancillary Service Auction Markets

239

such as supply offer caps and reliability unit commitments, could become less important if that takes place. Finally, this chapter has focused on the short-term, daily markets. As discussed in other chapters of this book, there is still evolution in the design of other elements of electricity market design to support investment decisions, such as resource adequacy or capacity markets and long-term ﬁnancial transmission rights. Because of the interactions between these different design elements, all aspects of the ISO market designs must be carefully integrated and calibrated [see, e.g., Stoft (2002) and O’Neill et al. (2006) for discussion].

5A. Appendix: Mathematical Formulation of the Auction Examples The auction examples given in several sections of this chapter employ a simpliﬁed version of the ISO auction designs, but with many of the main features of those markets, including realistic network power ﬂows including congestion and losses and including both start-up and energy offers by generators. In this appendix, the mathematical formulation of the auction model used in the examples is given, with some additional explanation. There are different ways to write this auction mathematically; the approach taken here is intended to improve the intuition of the model. For a more detailed mathematical description of the ISO auction on transmission networks, see, e.g., O’Neill et al. (2002) and the ISO auction manuals. A single period version of the auction model is as follows. Max

mi BIDmi dmi −

subject to

−yi +

EBL

mih STARTmih zmih + OFFERmih gmih

mh gmih − yi

fk+

≤

Fk+ max

=

m dmi

∀k k−

(3)

+ + + + − − − − k Dik fk − fk + fk Lki fk + fk Lki fk +

(1) (2)

∀k

fk− ≤ Fk− max

∀i i

k+ ≤ 0

∀i

−

Rf − f = 0 gmih ≤ Gmih zmih

∀m i h

(4) (5) (6)

dmi gmih fk+ fk− ≥ 0 zmih ∈ 0 1 where the notation is deﬁned as follows. Index sets H is the set of generators and virtual offers, h = 1 nH . I is the set of buses, i = 1 nI , in the transmission system. K is the set of transmission facilities, k = 1 nK . M is the set of participants in the auction market, m = 1 nM , whether as sellers or buyers (physical and virtual). Variables dmi is the quantity of energy bought by market participant m at bus i.

240

Competitive Electricity Markets

fk+ fk− represent the ﬂow on transmission element k in the positive and negative direction respectively (deﬁned arbitrarily). The net real power ﬂow on transmission element k could thus be deﬁned as fk , where fk = fk+ − fk− . f + f − are the vectors of these ﬂows. gmih is the quantity of energy, physical, or virtual, sold by market participant m at bus i from generator or virtual offer h. yi is the amount of real power injected at node i (withdrawn at node i if yi < 0) that is induced by the d bids and g offers that were awarded through the auction. zmih is the commitment decision associated with the offer for generator h submitted by market participant m at bus i. It is either 0 (uncommitted) or 1 (committed). In a dynamic model, separate variables would be deﬁned for the start-up decision (1 signaling a start-up occurs in a given hour) and commitment (1 indicating that the generator is operating in that period). i k+ k− are Lagrange multipliers associated with selected sets of primal constraints in the auction. Parameters and operators BIDmi is the bid price ($/MWh) submitted by market participant m at node i associated with a demand quantity dmi . In this context, “bid” means a bid to buy energy. OFFERmih is the offer price ($/MWh) submitted by market participant m for generator or virtual supply h at node i associated with quantity gmih . In this context, “offer” means a offer to sell energy. STARTmih is the start-up price ($) submitted by market participant m for generator h at node i. For all physical generators, this quantity is non-negative. Note that for all virtual offers, it is exactly zero. D is the arc incidence matrix, {Dki }. Dki = 1 if fk+ − fk− represents a MW ﬂow out of bus i through transmission line k in a positive direction; Dki = −1 if the ﬂow through k is in a negative direction; and Dki = 0 otherwise. Fk+ max Fk− max are transmission capacity constraints – thermal, stability, or contingency limits – associated with a transmission element k in the positive and negative directions. Gmih is the upper bound on the capacity offered by market participant m for generator or virtual supply offer h at node i. + L− ki Lki represent resistance loss coefﬁcients (decrease in imports to bus i) due to a negative and positive ﬂow, respectively, through transmission line k. R = rvk are line reactances used in Kirchhoff’s Voltage Law analogues. rvk is the value of reactance for transmission line k that appears in voltage loop v. rvk = +Rk or −Rk if line k occurs in loop v, depending on whether a positive fk+ − fk− is in the same or opposite sense of ﬂow around v. rvk = 0 if link k does not occur in loop v. Consistent with the linearized DC model of load ﬂow (Schweppe et al., 1988), the number of independent loops v must be equal to K − N + 1, where K is the number of lines considered and N is the number of buses.

EBL

The objective function maximizes social welfare, deﬁned as the sum of consumer surplus and producer surplus. This is the same as the integral of the demand curve (sum of accepted demand bids) minus as-bid production costs. Production costs include commitment costs, which are incurred if a generator is operating (i.e., zmih = 1), along with variable generation costs. More general versions include start-up and min run costs as separate terms in the objective. Constraint (1) is a net energy balance requirement for each

Design of US Wholesale Energy and Ancillary Service Auction Markets

241

bus on the network, whose dual variable is the LMP. Equations (2) and (3) are transmission capacity constraints on each transmission element. Their dual variables are the so-called “ﬂowgate prices.” Equations (4) and (5) are DC analogues to Kirchhoff’s Current and Voltage Laws. Equation (6) is the generation upper operating limit; if the unit is not committed (i.e., zmih = 0), then this constraint forces MW generation to be zero. Lower operating limits (“min run constraints”) and ramp rate constraints are not shown, but could be introduced. The numerical example takes place on the three-bus network in Fig. 5.1, in which the arrows show the direction of ﬂow for an injection at bus 1 and a withdrawal at bus 3 (note that the arrows do not correspond to the direction of the ﬂowgates). In the equations that follow, the three transmission lines indexed k above are labeled for the two buses to which they are connected in the “positive” direction. Hence, the line from bus 1 to bus 2 is labeled “12,” the line from bus 1 to 3 is labeled “13,” and the line from bus 2 to 3 is labeled “23.” All loss factors on all lines = 0.00001 [MW/MW2 ]. All reactances, Rk = 1. Then (4) for each bus becomes + + − − − 2 − 2 KCL1 −y1 + f12 − f12 + f13 − f13 + 00001f12 + 00001f13 ≤ 0 + + + 2 − − − 2 − f12 + f23 − f23 + 00001f12 + 00001f23 ≤ 0 KCL2 −y2 − f12 + + − − − 2 − 2 − f23 − f13 − f13 + 00001f23 + 00001f13 ≤ 0 KCL3 −y3 − f23

and (5) becomes + + + − − − KVL f12 − f12 + f23 − f23 − f13 − f13 = 0

EBL

With these simpliﬁcations, the auction examples can be replicated using commercially available software, such as GAMS. Acknowledgment The authors would like to thank David Mead, Partha Malvadkar, and Harry Singh for useful comments, and Perry Sioshansi for editorial advice and great patience. References44 California ISO (2005). Locational Marginal Pricing (LMP) Study 3C, Analysis of Market-Based Price Differentials, Description of Methodology, Appendix A, Disaggregation of LMP Components. 19 November. Available at http://www.caiso.com/14cd/14cd6d735cf60.pdf. Cramton, P. and Stoft, S. (2006). Uniform-Price auctions in electricity markets. Working Paper, University of Maryland, 20 March. Fabra, N., von der Fehr, N., and Harbord, D. (2004). Designing electricity auctions. Center for the Study of Energy Markets, University of California, Berkeley, CSEM WP-122, February. Federal Energy Regulatory Commission (FERC) (1996a). Promoting Wholesale Competition Through Open Access Non-Discriminatory Transmission Services by Public Utilities; Recovery of Stranded Costs by Public Utilities and Transmitting Utilities. Order No. 888, FERC Stats. & Regs. 31,036, 24 April.

44 The technical manuals referenced below were all current as this chapter was being written. Such manuals are periodically updated to reﬂect market rule changes. The websites of the ISOs whose materials are cited are as follows: PJM, www.pjm.com; New York ISO, www.nyiso.org; ISO-New England, www.iso-ne.com; Midwest ISO, www.midwest.org

242

Competitive Electricity Markets

Federal Energy Regulatory Commission (FERC) (1996b). Open Access Same-Time Information System (formerly Real-Time Information Networks) and Standards of Conduct. Order No. 889, 75 FERC 61,078, 24 April. Federal Energy Regulatory Commission (FERC) (1997). PJM Interconnection, L.L.C., et al., Order Conditionally Accepting Open Access Transmission Tariff and Power Pool Agreements, Conditionally Authorizing Establishment of An Independent System Operator and Disposition of Control over Jurisdictional Facilities, and Denying Rehearing. 81 FERC 61,257, 25 November. Federal Energy Regulatory Commission (FERC) (1999a). Central Hudson Gas & Electric Corporation et al., Order Conditionally Accepting Tariff and Market Rules, Approving Market-Based Rates, and Establishing Hearing and Settlement Judge Procedures, 86 FERC 61,062, 27 January. Federal Energy Regulatory Commission (FERC) (1999b). Regional Transmission Organizations. Order No. 2000, 89 FERC 61,285, 20 December. Federal Energy Regulatory Commission (FERC) (2000). PJM Interconnection, L.L.C., Order Accepting Tariff Sheets as Clariﬁed. 92 FERC 61, 013, 7 July. Federal Energy Regulatory Commission (FERC) (2002). Standard Market Design and Structure, Notice of Proposed Rulemaking. 18 CFR Part 35, Docket Number RM01-12-000, September. Federal Energy Regulatory Commission (FERC) (2005). Principles for Efﬁcient and Reliable Reactive Power Supply and Consumption. FERC Staff Report, AD05-1-000, 4 February. Guan, X., Zhai, Q., and Papalexopoulos, A. (2003). Optimization based methods for unit commitment: Lagrangian relaxation versus general mixed integer programming. Power Engineering Society General Meeting, IEEE. Harvey, S.M., Hogan, W.W., and Pope, S. (1997). Transmission Capacity Reservations and Transmission Congestion Contracts. John F. Kennedy School of Government, Harvard University, Cambridge, MA. June 6, 1996. revised 8 March. Helman, U. (2006). Market power monitoring and mitigation in the US wholesale power markets. Energy, 31(6–7), 877–904. Hobbs B.F., Rothkopf, M., Hyde, L., and O’Neill, R.P. (2000). Evaluation of a truthful revelation auction in the context of energy markets with nonconcave beneﬁts. J. Reg. Econ., 18(1), 5–32. Hobbs, B.F., Stewart, W.R. Jr., Bixby, R.E., et al. (2001). Why this book?: New capabilities and new needs for unit commitment modeling. In The Next Generation of Electric Power Unit Commitment Models (B.F. Hobbs, M.H. Rothkopf, R.P. O’Neill, and H-P. Chao, eds). Boston: Kluwer Academic Press, pp. 1–14. Hogan, W.W. (1993). Markets in real electric networks require reactive prices. Energ. J., 14(3), 171–200. Hogan, W.W. (2006). Resource sufﬁciency guarantees and cost allocation. Comments submitted to the Federal Energy Regulatory Commission, Docket No. ER04-691-065, 25 May. Available at http://ksghome.harvard.edu/˜whogan/. Johnson, R.B., Oren, S.S., and Svoboda, A.J. (1997). Equity and efﬁciency of unit commitment in competitive electricity markets. Util. Pol., 6, 9. Kahn, A.P., Cramton, P., Porter, R., and Tabors, R. (2001). Pricing in the California Power Exchange electricity market: Should California switch from uniform pricing to pay-as-bid pricing? Blue Ribbon Panel Report, Prepared for the California Power Exchange, January. Kahn, E. and Baldick, R. (1994). Reactive power is a cheap constraint. Energ. J., 15(4), 191–202. Klemperer, P. (1999). Auction theory: A guide to the literature. J. of Econ. Surv., 13(3), 227–86. Krishna, V. (2002). Auction Theory. San Diego, Calif.: Academic Press. New York ISO (1999). Transmission and Dispatching Operations Manual, Manual 12, September. New York ISO (2001a). Market Services Tariff. Issued on 16 January. New York ISO (2001b). Day-Ahead Scheduling Manual, Manual 11, June. New York ISO (2004). Subject: Scheduling a “Must-Run” Generator, Technical Bulletin 026, 10 November. New York ISO (2005). Subject: Allocation of Uplift Costs to Load and Other Entities Associated with Virtual Trading, Technical Bulletin 082, 5 November, 2001; updated 4 November, 2005. New York ISO (2006). Ancillary Services Manual, Manual 2, May. O’Neill, R.P., Baldick, R., Helman, U., Rothkopf, M.H., and Stewart, W.R. (2005). Dispatchable transmission in RTO markets. IEEE Trans. on Pow. Sys., 20(1), 171–9.

EBL

Design of US Wholesale Energy and Ancillary Service Auction Markets

243

O’Neill, R.P. and Helman, U. (2007). Regulatory reform of the U.S. wholesale electricity markets. In Creating Competitive Markets: The Politics of Regulatory Reform (M.K. Landy, M.A. Levin, and M. Shapiro, eds). Brookings Press. O’Neill, R., Helman, U., Hobbs, B.F., and Baldick, R. (2006). Independent system operators in the USA: History, lessons learned, and prospects. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. O’Neill, R.P., Helman, U., Hobbs, B.F., et al. (2002). A joint energy and transmission rights auction: proposal and properties. IEEE Trans. Pow. Sys., 17(4), 1058–67. O’Neill, R.P., Sotkiewicz, P.M., Hobbs, B.F., et al. (2005). Efﬁcient market-clearing prices in markets with nonconvexities. Eur. J. Op. Res., 164, 269–85. PJM (2000). State of the Market Report, 1999. Market Monitoring Unit, PJM Interconnection, L.L.C., June. PJM (2001). State of the Market Report, 2000. Market Monitoring Unit, PJM Interconnection, L.L.C., June. PJM (2002). State of the Market Report, 2001. Market Monitoring Unit, PJM Interconnection, L.L.C., June. PJM (2003). State of the Market Report, 2002. Market Monitoring Unit, PJM Interconnection, L.L.C., 5 March. PJM (2004). State of the Market Report, 2003. Market Monitoring Unit, PJM Interconnection, L.L.C., 4 March. PJM (2005). State of the Market Report, 2004. Market Monitoring Unit, PJM Interconnection, L.L.C., 8 March. PJM (2006a). State of the Market Report, 2005. Market Monitoring Unit, PJM Interconnection, L.L.C., 8 March. PJM (2006b) PJM Manual 15: Cost Development Guidelines. PJM Interconnection, L.L.C., Revision 07, Effective Date: 3 August. PJM (2007a). State of the Market Report, 2006. Market Monitoring Unit, PJM Interconnection, L.L.C., 8 March. PJM (2007b). PJM Manual 11: Scheduling Operations. PJM Interconnection, L.L.C., Revision 30, Effective Date: 20 March. PJM (2007c). PJM Manual 6: Financial Transmission Rights. PJM Interconnection, L.L.C., Revision 9, Effective Date: 4 April. PJM (2007d). PJM Manual 12: Balancing Operations. PJM Interconnection, L.L.C., Revision 15, Effective Date: 15 May. PJM News (2004). PJM’S New Generation-Scheduling Software to Save Customers Estimated $56 Million: Unit Commitment System Uses First of its Kind Technique. Valley Forge, Pa., 24 June. Potomac Economics, Ltd. (2003). 2002 State of the Market Report, New York ISO. Independent Market Advisor to the New York ISO, June. Potomac Economics, Ltd. (2004). 2003 State of the Market Report, New York ISO. Independent Market Advisor to the New York ISO, May. Potomac Economics, Ltd. (2005). 2004 State of the Market Report, New York ISO. Independent Market Advisor to the New York ISO, July. Potomac Economics, Ltd. (2006). 2005 State of the Market Report, New York ISO. Independent Market Advisor to the New York ISO, August. Schweppe, F. C., Caramanis, M. C., Tabors, R. D., and Bohn, R.E. (1988). Spot Pricing of Electricity. Boston: Kluwer Academic Publishers. Stoft, S. (2002). Power System Economics. New York: IEEE Press. Streiffert, D., Philbrick, R., and Ott, A. (2005). A mixed integer programming solution for market clearing and reliability analysis I. In Power Engineering Society General Meeting, 2005, IEEE, San Francisco, CA. Sweeney, J.L. (2006). California electricity restructuring, the crisis and its aftermath. In Electricity Market Reform: An International Perspective (F. Sioshansi and W. Pfaffenberger, eds). Amserdam: Elsevier. Wilson, R. (2002). Architecture of power markets. Econometrica, 70(4), 1299–340.

EBL

EBL

This page intentionally left blank

Chapter 6 The Cost of Anarchy in Self-Commitment-Based Electricity Markets RAMTEEN SIOSHANSI1 , SHMUEL OREN1 , AND RICHARD O’NEILL2 University of California, Berkeley, USA; 2 Federal Energy Regulatory Commission, USA

1

Summary

EBL

With the advent of restructured electricity markets a contentious market design issue has been whether unit commitment decisions should be made centrally by the system operator or by individual generators. Although a centrally committed market can, in theory, determine the most efﬁcient commitment, they have been shown to suffer some equity and incentive problems. A self-committed market can overcome some of these incentive issues, but will generally suffer efﬁciency losses from not properly coordinating commitment and dispatch decisions between individual generators. This chapter examines the issue of dispatch efﬁciency raised by the design of markets based on central versus self-commitment by determining a set of “competitive benchmarks” for the two market designs. Comparing the total commitment and dispatch costs of the two markets provides a bound on the productive efﬁciency losses of a self-committed market. 6.1. Introduction The introduction of competition in the electric supply industry has led to a number of important questions regarding the need for organized markets to efﬁciently and reliably coordinate the power system and the desirable features and scope of those markets. Complicating electricity market design, power systems are subject to a number of “network” constraints, in that these constraints depend on the actions of every market participant and each participant can impose an externality on others using the power system. That is, the ability of customer A to purchase power from generator B can depend on the actions of generator C or of consumer D. These complexities have called into question the ability of decentralized markets to efﬁciently and feasibly commit and dispatch units 245

246

Competitive Electricity Markets

while respecting power system constraints and the market design rules for any centralized markets operated by a system operator (SO). While a centralized market can, in theory, ﬁnd the most efﬁcient dispatch of the generators, the market designs suffer equity and incentive problems. Decentralized designs can overcome some of these issues but will suffer efﬁciency losses due to the loss of coordination between market participants. As electricity markets in various countries have been restructured and evolved, different approaches have been used with varying degrees of success. In the United States, e.g., the move toward standard market design has led to heavy reliance on open and transparent centralized markets. The British market, by contrast, started with a centralized market in the original Electricity Pool and moved to a more decentralized design under the New Electricity Trading Arrangements (NETA) and subsequent reforms under the British Electricity Trading and Transmission Agreements (BETTA), which were meant to overcome some of the problems experienced under the original centralized pool design. Although these design differences are driven by realities of the market such as asset ownership, generation mix, and system infrastructure, different “philosophies” regarding the proper role of centralized markets have also played a role in determining the scope of any centralized markets. Two of the chapters in this volume, Chapter 5 and Chapter 11, e.g., advocate centralized versus decentralized markets, respectively, based on these philosophies and experiences in different markets. This chapter revisits the issue of dispatch efﬁciency raised by the design of markets based on centralized versus decentralized dispatch. Using actual market data from an SO, a set of “competitive benchmarks” for a centralized and decentralized market are computed and compared. These competitive benchmarks assume that generators do not behave strategically in manipulating their offers in the two markets. Comparing the total commitment and dispatch cost of the two market designs shows the extent of productive efﬁciency losses from a decentralized market, which are small but non-trivial. Comparing total settlements costs of the two designs shows that generators can extract signiﬁcantly higher payoffs from consumers under a decentralized design, suggesting that this design may be the wrong approach if the goal of restructuring is to reduce consumer costs. These higher prices under a decentralized design would further result in allocative efﬁciency losses in markets with demand response.1 The remainder of this chapter is organized as follows. Section 6.2 provides the context on the debate regarding the scope of centralized markets. Sections 6.3.1 and 6.3.2 further describe centrally and self-committed markets and the market models used in the simulations, while Section 6.3.3 provides the results and comparison of the two designs. Section 6.4 provides conclusions. The Appendix 6A outlines the formulations and algorithms used in the simulations and analysis.

EBL

6.2. Centralized versus Decentralized Markets Unlike other commodities, electricity has a number of technical constraints, which must be obeyed to ensure feasible and reliable service. Complicating electricity market design, power systems are subject to physical constraints, which make markets for electricity inherently incomplete. This incompleteness is due in part to costly storage of electricity, 1 As in any market with price-elastic demand, prices which are above the competitive level will result in higher producer proﬁts at the cost of consumer welfare losses.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

247

complex power ﬂows within the transmission network, extremely price-inelastic short-run demand for energy, and the need to constantly balance supply and load in real-time. Injections and withdrawals of electricity cannot be directed along a speciﬁed path within a transmission network, rather power ﬂows along each line in inverse proportion to the line’s impedance. Flows on lines within the network are constrained by physical limitations and environmental factors. Moreover, allowing an injection of power, which will ﬂow in a constrained (also known as congested) direction along a path, may be contingent upon another injection, which provides “counterﬂows” to relieve congestion on that path. These characteristics of power systems mean that use of the transmission network for injection and withdrawal of energy must be properly coordinated, otherwise the resulting dispatch may not be simultaneously feasible or optimal. Generating units are constrained in the time it takes them to start up or shut down, and the rate at which they can adjust their output. Thermal units have non-zero minimum generating constraints and “forbidden zones,” in which they cannot operate stably when they are online. Other types of generating units, such as combined-cycle gas turbines (CCGT) and cascaded watershed hydroelectric systems, tend to have complex constraints restricting their operation. Due to the stochastic nature of demand ﬂuctuations, generators must be able to adjust their real and reactive power outputs in real-time to ensure constant load balance. Other random contingencies such as transmission equipment failures or forced generator outages may also require generators to adjust their outputs within a short period of time to maintain system reliability. Thus, efﬁcient and reliable operation of the system requires having a sufﬁcient number of generators online and available to react to variations in load and other contingencies at least cost. These physical realities of power systems have given rise to a debate about whether competitive pressures in a decentralized or bilateral market can lead to efﬁciency and reliability of the system, or if we must rely on centrally organized markets. Whatever market design is ultimately adopted, there are several objectives it should address, including the following:2

EBL

•

Efﬁcient dispatch: The market should ensure that generators within the system are efﬁciently dispatched to meet system load at least cost, while satisfying the requisite operational constraints. • Network feasibility: The pattern of injections and withdrawals within the transmission network and the resulting power ﬂows must be feasible. • System reliability: The market must ensure that there is sufﬁcient excess generating capacity online and available (known as ancillary services) to react to contingencies such as load ﬂuctuations, or generator or transmission outages within a sufﬁciently short amount of time to ensure reliability. This market design debate has often been framed as being between the two extremes of the bilateral versus Poolco model, as described by Hogan (1994). The bilateral model emphasizes direct transactions between buyers and sellers. It argues that competitive forces in the marketplace will resolve any infeasibilities in dispatches and schedules and deliver efﬁcient and reliable electric service, without the need to design or impose new market institutions. Technical difﬁculties such as load balance, power ﬂow constraints, and ancillary services would be resolved by market transactions. According 2 Joskow and Schmalensee (1983) and Hunt (2002) provide a more thorough discussion of important considerations in electricity market design.

248

Competitive Electricity Markets

to proponents of the bilateral model, competitive forces would minimize costs while capturing any potential efﬁciency gains, desirable features that may be hampered by the rigidities of centralized markets. The Poolco model, on the other hand, emphasizes the need for tight coordination of the power system to ensure efﬁcient, feasible, and reliable service. This model has inﬂuenced the Standard Market Design (SMD), “blessed” by the Federal Regulatory Energy Commission in the US, and variants of this model have been implemented (or implementation is underway) in about 50% of US states. Proponents of the Poolco model have argued that power balance, provision of ancillary services, and feasible power ﬂows and counterﬂows within the transmission network may not be easily achieved through bilateral transactions. Moreover, the system must be in continuous real-time balance to ensure reliable operation, meaning the bilateral model may have to rely on a series of short-run minute-by-minute bilateral transactions to constantly adjust schedules and ensure system feasibility. Hogan (1994), e.g., paints an extreme picture of a “strict bilateral market,” in which the market is completely reliant on bilateral transactions without any centralized markets. His example envisions that consumers3 must contract with generators to provide energy and other essential services, and manage these contracts in real-time, independently of one another. In his example, each customer and generator would have to track changes in one another’s output and consumption and make adjustments to their own generation and load or make spot bilateral transactions to correct load imbalances. Furthermore, the two parties would have to adjust their portfolio of transmission capacity rights to enable the actual real-time power ﬂows. Much of this debate originally centered on use of the transmission system. Ruff (1994) advocates development of organized spot markets, in which energy and transmission are traded in an integrated fashion, as opposed to relying on bilateral trading between the incumbent utilities and new entrants. He argues that in the absence of organized spot markets the industry may fall back on traditional wheeling agreements, which are ad hoc trades that were used by vertically integrated utilities under the regulated industry paradigm to handle a few incremental trades with other utilities. These wheeling agreements would be, in his opinion, inadequate to deliver all the potential efﬁciency gains and cost savings from trading in a restructured wholesale market. Hogan (1995) further advocates a centralized Poolco model. He argues that the physical realities of power ﬂows would make the allocation of physical transmission rights to parties difﬁcult, and as such a bilateral approach may not fully realize efﬁcient trade. In his opinion, only a centralized market which cooptimizes generation and transmission ﬂows will achieve least-cost dispatch. Although much of the debate initially focused on use of the transmission network, another important facet of this market design issue is how the commitment (on/off) status of generators should be determined. Under the traditional regulated monopoly paradigm, utilities would determine unit commitments to minimize total costs while meeting load, ancillary service, transmission, and generator operational constraints over a ﬁxed planning period, typically between 24 and 168 hours. One of the difﬁculties that has plagued the solution of unit commitment problems is their computational complexity. Because commitment decisions are binary (i.e., a generator is either on or off in a given hour),

EBL

3 The term “consumer” is used throughout this chapter to encompass any entity purchasing in the wholesale market. This includes both large customers purchasing electricity for themselves and loadserving entities.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

249

they must be formulated as mixed-integer programs (MIPs), which are computationally demanding, as shown by Guan et al. (2003) and others. As such, utilities have had to rely on approximation algorithms such as Lagrangian relaxation (LR) to ﬁnd a near- (but sub-) optimal feasible solution within a reasonable amount of time, so the utility can follow the least-cost commitment.4 In a competitive setting, the market must provide for this unit commitment planning. The Poolco model typically envisions unit commitments being determined in a centralized manner by the SO. The SO solicits generation offers from generators, which specify their operating constraints and cost structure,5 and combines these with hourly load forecasts to determine a least-cost commitment and dispatch, which satisﬁes the requisite generator and system constraints. Ancillary service constraints are normally based on possible generator and transmission contingencies. For instance, it is typical to ensure that there is sufﬁcient excess generating capacity available to react to the forced outage of the largest and second largest generators within a set number of minutes. This unit commitment process is typically done day-ahead, meaning commitments are determined for the following day and an advisory schedule is given to each generator specifying their generation in each hour.6 Once the commitments are ﬁxed, on the following day, the SO will redispatch generators in real-time to meet the actual load and obey real-time system constraints. This allows the dispatch to react to contingencies such as load shocks, forced generator outages, or changes in transmission network topology. Because most generators incur ﬁxed costs when starting up and when online, the cost offers generators submit to the SO typically consist of three parts:7 • •

EBL

A start-up cost, which is incurred whenever the unit is started from an ofﬂine state. A no-load cost, which is the spinning cost a generator incurs from being online, regardless of its actual generation. • Its actual variable generation costs, which are typically expressed as a nondecreasing step- or piecewise linear-function. The bilateral model, by contrast, leaves these commitment decisions to individual generators. In a purely bilateral market, generators will contract to deliver energy to consumers. This contracted energy will then, in real-time, be supplied by some combination of producing energy from its own generating assets or contracting with another party to deliver the energy. To produce energy itself, a generator will have to determine the commitment status of its own units to ensure it can deliver the energy. Similarly, consumers can contract with generators to hold excess generating capacity for ancillary services to ensure its demand can be reliably served under different system conditions, which will again 4

Muckstadt and Koenig (1977) provided one of the ﬁrst formulations of unit commitment problems and proposed the use of the LR algorithm in their solution. Hobbs et al. (2001) and O’Neill et al. (Chapter 5) provide more recent treatments of unit commitment. 5 Although some authors use the term “bid” to describe the cost and operating constraint data provided to the SO by a generator, this chapter calls them “offers,” which distinguishes a generator’s offer to generate energy from a consumer’s bid to purchase energy. 6 Some markets with centralized unit commitment have longer planning horizons. The new market design put forth by the California ISO, e.g., has a separate unit commitment for the second day out, which is used to advise units with long start times to have themselves ready to be online. 7 Some SOs also allow generators to specify a shutdown cost, which is incurred whenever the unit is halted from an online state. Generators operating in markets which do not accept shutdown costs can roll these shutdown costs into the unit’s start-up cost.

250

Competitive Electricity Markets

factor into a generator’s unit commitment decisions. If generators are liable for damages from not serving a customer’s load, the generator may include its own ancillary service requirements in its commitment decision. Alternately, some bilateral designs call for the SO or another entity to operate energy and ancillary service markets into which generators can offer their generation. These markets are typically operated day-ahead, and in many cases hour-ahead or markets with other timeframes are operated. Although the assignment of generators to generate or provide ancillary services is done in a centralized manner in such a market, generators individually decide whether to commit themselves in expectation of revenues from the market. Wilson (1997), for instance, suggested the use of such a market in the original restructured California market. In a strictly bilateral market, real-time load imbalances would have to be resolved directly between contract counterparties, for instance, as described by Hogan (1994). Otherwise, some bilateral markets have real-time balancing service markets in which imbalance energy (both incremental and decremental) can be traded. Besides philosophical differences regarding the extent to which transactions in the market should be standardized and centralized within the SO’s purview, there are also incentive and efﬁciency issues arising from the two designs. The clear advantage of having the SO make commitment decisions in a centralized manner is that a centralized market will, in theory, ﬁnd the most efﬁcient commitment and dispatch of units. Generators acting independently of one another may not be able to achieve this, because the non-convex nature of generator operating constraints and costs mean there are efﬁciencies to be gained from coordination amongst generators.8 Moreover, a centrally committed market could easily ﬁnd a feasible schedule, while this may prove more difﬁcult in a decentralized paradigm. Centrally committed markets do, however, suffer from incentive issues, which call into question the efﬁciency of the underlying commitment. For one, the auction into which generators submit their operating constraints and costs is not strategy-proof, meaning generators may ﬁnd it proﬁtable to misstate their parameters. Oren and Ross (2005) demonstrate, by studying simple examples, that generators can proﬁtably misstate their ramping constraints. Because these parameters are then used by the SO in determining the commitment of units, this type of misrepresentation on the part of generators can result in productive efﬁciency losses. That is, the SO could ﬁnd a commitment which appears to be least-cost on the basis of the parameters given to it by generators, but which is sub-optimal because these parameters are misstated. Furthermore, Johnson et al. (1997) highlight other equity, incentive, and efﬁciency issues with centralized unit commitment. They demonstrate that if the SO solves for a commitment using an LR algorithm and makes binding commitment and dispatch decisions on the basis of this solution, different near-optimal solutions which are the same in terms of total commitment costs can result in vastly different payoffs to individual generators. Thus, seemingly benign programming parameters and heuristics used in the LR algorithm could arbitrarily determine which generators are “winners” and “losers” in being committed and receiving energy payments. Generators, knowing the nature of the LR commitments, may be further inclined to misstate their constraint and cost parameters to increase their likelihood of being committed and dispatched, which would again

EBL

8 Non-convexities imply that an efﬁcient Walrasian equilibrium with linear prices may not exist. See O’Neill et al. (2005) for a more detailed discussion.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

251

call into question the efﬁciency of the solution due to incorrect data in the underlying problem. Although recent advances in computational capabilities and optimization software now make it tractable to solve the MIP formulation of a commercial-scale SO unit commitment problem, Sioshansi et al. (2007) show that eliminating the issues with LR solutions raised by Johnson et al. (1997) requires solving the MIP to complete (as opposed to near-) optimality. They demonstrate that otherwise, if the algorithm used to solve the MIP must be stopped by the SO prematurely, the sub-optimal solution can yield generator payoffs which are different from the optimum. They further demonstrate that because energy prices are determined by the marginal cost of generation in each hour, which is in turn determined by the set of units committed in each hour, these energy prices can deviate signiﬁcantly from the optimum as well. Moreover, they show that the relative size of these payoff and energy price deviations do not necessarily decrease as the commitment gets closer to optimal – meaning the only way to completely eliminate the issues raised by Johnson et al. (1997) is to solve the MIP to complete optimality. Although some SOs with central commitment have adopted MIP (as opposed to LR) in the solution of their unit commitment problems, none of these SOs are able to solve their unit commitment to complete optimality within the limited timeframe they have to determine the next day’s commitments and advise generators, implying the issues raised by Johnson et al. (1997) have not been alleviated in these markets. The issues raised by Johnson et al. (1997) have been used to champion simpler market designs in which unit commitment decisions are made in a decentralized fashion by individual generators. Advocates of decentralized commitment claim it will reduce incentive compatibility issues since generators must internalize their operating constraints while minimizing production costs. Furthermore, if generators are making commitment decisions to maximize proﬁts in expectation of energy prices, this process is similar to solving a centralized unit commitment by means of an LR algorithm.9 Wilson (1997) proposed a market design with self-commitment and a multi-round energy auction in which generators and loads would iteratively submit one-part offers specifying prices at which they would be willing to supply and consume energy. The offers would be iteratively updated, subject to some proposed activity rules,10 until the prices and dispatch converged. His proposal was geared toward the original California market design, and while

EBL

9 When an LR algorithm is used to solve a central SO unit commitment problem, it works by relaxing load balance and ancillary service constraints and penalizing these in the objective function. The problem then decouples in the sense that there are no longer any constraints linking the decision variables of the individual generators, and each can be solved independently of the others. The algorithm then works to ﬁnd coefﬁcients for the penalty terms which incent sufﬁcient generation and reserves to satisfy the relaxed constraints. These coefﬁcients can, in turn, be thought of as energy and ancillary service prices, and each generator’s decoupled problem can be rewritten as a proﬁtmaximization problem, which would be equivalent to a market with decentralized commitment. The important distinction, however, is that because a centralized unit commitment problem typically has a duality gap, heuristics must be used to ensure primal-feasibility of the dual solution. This heuristic process can commit a unit at a net proﬁt loss, meaning that the ﬁnal feasible centralized solution can differ from the decentralized in terms of some near-marginal commitments and dispatches. See Wolsey (1998) for a further discussion of the LR algorithm. 10 The purpose of the activity rules is to ensure early price discovery, fast convergence, and to prevent large generators and consumers from manipulating the auction by withholding themselves from the market until the ﬁnal round of bidding.

252

Competitive Electricity Markets

some of the elements were used, the design eventually settled on a simpler single-round auction. Given these divergent philosophies on the fundamentals of market designs and the pros and cons of centralized versus decentralized markets, different jurisdictions have implemented a variety of designs spanning the spectrum between the bilateral and Poolco models. Britain, which was one of the ﬁrst markets to restructure, originally established the centralized Electricity Pool, which was a mandatory market through which all energy was transacted.11 The pool operated as a day-ahead market which accepted generation offers consisting of three-part costs and unit operation constraints, and determined a least-cost commitment and dispatch of the system. The pool relied on the same software program, GOAL, which was used by the vertically integrated monopolies prior to restructuring to determine these commitments. Because generators had access to the GOAL software, they were able to manipulate and misstate the cost and constraint parameters in their offers in such a way to maximize proﬁts, as opposed to submitting the true values. A review of the pool, conducted by Offer (1998), further criticized the GOAL software as being opaque in its pricing (due to various heuristics employed in the feasibility phase of the LR algorithm12 ) and using scheduling algorithms that are more appropriate for a vertically integrated utility than for a competitive market. The British market was eventually restructured under NETA, which established a series of overlapping voluntary markets, that rely primarily on bilateral and long-term contracting between generators and consumers. NETA has a minimal balancing market based on a pay-as-bid auction three and a half hours prior to physical delivery. NETA was further reformed under BETTA, but still maintained a decentralized design with self-commitment. The market in Texas, established in 2001, is based almost entirely on individual bilateral transactions that rely on generators to commit their units individually.13 The bulk of energy and ancillary services (typically 95–98%) are traded and contracted on a long-term basis between generators and consumers, with the SO operating markets for imbalance energy and to acquire ancillary services as a provider of last resort. Day-ahead, generators, and consumers submit schedules to the SO, specifying expected injections and withdrawals of energy and ancillary service schedules. Originally, market rules required consumers to submit balanced schedules – meaning scheduled injections had to cover expected loads. This balanced schedule requirement was meant to encourage consumers to contract for their expected load as opposed to relying on the balancing market for procurement of base load energy. Eventually, in mid-2002, the balanced schedule requirement was dropped, yet the balancing market still accounted for only a small fraction of energy trades.14 Due to the inaccuracy of load forecasts and because of random events, such as generator or transmission outages, generators and consumers would have to have a means of

EBL

11

Newbery (2006) provides a more detailed description of the history and structure of the British electricity market. 12 A Branch and Bound approach, on the other hand, guarantees feasibility of the solution but nontheless would require some heuristics to decide when to terminate the search since running the algorithm to completion is typically unrealistic in a practical setting. 13 The Texas market is currently undergoing a major reform which among other things will establish a voluntary day-ahead market with centralized unit commitment-based three-part offers. See Adib and Zarnikau (2006) for a more detailed discussion of the Texas market. 14 See Hortaçsu and Puller (2005) for a discussion of the implications of the balanced schedule requirement.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

253

making real-time adjustments to their schedules. As opposed to requiring generators and consumers to make these adjustments individually, the SO operates a balancing energy services (BES) market. Generators can submit offers for both incremental and decremental energy, which is then dispatched in real-time to ensure load balance, but the offers are simple one-part energy-only offers specifying a price at which a generator is willing to increase or decrease its generation from its schedule. The BES market is further used by the SO to purchase imbalance energy in order to relieve congestion on transmission lines. Because of the highly decentralized nature of the Texas market (with the exception of the pool-like BES market), there is no centralized mechanism by which commitment decisions are made. Rather, generators individually determine commitments for their units in order to meet their contractual obligations and any imbalance or ancillary service15 sales they expect to make. The two examples presented are meant to be illustrative of centrally16 versus selfcommitted markets, with other markets falling between these two examples. Table 6.1 summarizes where other major restructured markets fall with regard to determining unit commitments. The Australian National Electricity Market, original California market, and NordPool are further examples of self-committed markets. The original California market is slightly unique in that it is a hybrid of a Poolco model with self-commitment. Generators would make commitment decisions individually, but nearly all energy was traded

Table 6.1. Examples of centralized versus decentralized unit commitment in selected electricity markets

EBL

Market Australian National Electricity Market Brazil British California Chile Columbia ISO New England Midwest ISO New York New Zealand NordPool PJM Texas ∗

Electricity Pool NETA/BETTA Original design MRTU∗ redesign

Current design Texas Nodal redesign

Commitment

Source

Self

Moran (2006)

Self Central Self Self Central Self Self Central Central Central Self Self Central Self Central

Dyner et al. (2006) Newbery (2006) Sweeney (2006) and California ISO (2007) Raineri (2006) Dyner et al. (2006) ISO New England (2007) Midwest ISO (2005) New York ISO (2001) Bertram (2006) Amundsen et al. (2006) Bowring (2006) Adib and Zarnikau (2006)

Market Reform and Technology Update

15

The ancillary service market operates similarly to the BES, into which generators submit offers specifying prices at which they are willing to be held in reserve for different qualities (i.e., reaction times) of ancillary services. 16 Most centrally committed markets do allow for self-commitment.

254

Competitive Electricity Markets

day-ahead in a mandatory central power pool. The three large incumbent utilities were initially prevented, by regulatory mandate, from contracting for energy on a long-term forward basis.17 This restriction was later relaxed but the utility had no incentive to contract since the regulator would not guarantee that the contract cost would be considered “prudent” if it turned out to be higher than the average spot price on the central power exchange. ISO New England, PJM, the proposed California market redesign, and the Texas market redesign are examples of centrally committed markets.

6.3. Comparison of Centrally and Self-Committed Markets Clearly, centrally and self-committed markets present tradeoffs, which must be evaluated in addressing market design issues. Centrally committed markets strive for the least-cost commitment and dispatch of generators by solving for a commitment which minimizes the SO’s cost objective. Theoretical studies and empirical observations have demonstrated incentive and equity issues with centralized unit commitment, which call into question the efﬁciency of the central solution, however. Self-commitment has been offered as a viable alternative, which addresses and reduces some of the issues with centrally committed markets. Self-commitment will, however, suffer from some loss of coordination amongst generators, resulting in efﬁciency losses. Thus, central and self-commitment are two imperfect market models with inherent shortcomings.18 To compare the two designs, this section describes and analyzes models of the two markets. The models assume that the markets are competitive – thereby eliminating the incentive issues and focusing instead on the relative efﬁciency losses and settlement costs of relying on self- as opposed to central commitment. Unit commitment refers to determining a short-term schedule of the on/off status of generating units to ensure sufﬁcient resources are available to serve load while satisfying transmission network and contingency constraints at least cost. In many restructured markets commitments are determined day-ahead (i.e., the day before the commitments are to take place) with a planning horizon of 24 single-hour periods. Some markets have different planning horizons or period lengths in their commitment process. The proposed market redesign for the California ISO will consist of a day-ahead unit commitment and a separate one-day commitment for the second day out. The purpose of this second unit commitment is to give advanced warning to units with extremely long start times that they will be needed to be online. The Electricity Pool in the original British market had a single-day planning horizon, but the commitment model consisted of 48 half-hour-long periods. The simulations and discussion in this section will assume the more standard planning horizon of 24 single-hour periods.

EBL

17

The one exception to this was a short-lived and highly illiquid block-forward market which the California Power Exchange (the entity which administered the day-ahead spot market) operated. 18 In theory, a Vickery–Clarke–Groves (VCG) mechanism (assuming the SO’s unit commitment problem could be solved to complete optimality) would address both the incentive and efﬁciency issues. This suffers the obvious shortcoming that the SO unit commitment cannot be solved to optimality. In fact, the VCG mechanism requires solving the unit commitment problem multiple times – once for each generation ﬁrm in the market. Moreover, VCG payments are discriminatory, complicated, and not budget balanced, making the mechanism an unrealistic option. See Mas-Colell, Whinston, and Green (1995) for a discussion of mechanism design and the VCG auction.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

255

6.3.1. Centrally committed model In centrally committed markets, the SO typically forecasts transmission network constraints for the following day, solicits demand bids from consumers,19 self-schedules,20 and virtual bids,21 and generation offers consisting of costs and operating constraints. For ease of problem formulation, an SO which centrally commits units accepts costs and operating constraints as a set of standardized parameters. Most SOs accepts three-part costs, consisting of start-up, no-load, and marginal generating costs. Start-up costs can usually be declared as being time-dependent with, for instance, a different start-up cost when a unit is in a hot, intermediate, or cold state, with the state being dependent on how long a unit has been ofﬂine. Marginal generating costs are normally submitted as non-decreasing stepwise- or piecewise-linear functions. Operating constraints are similarly given as a set of parameters. For example, a ramp rate indicating the maximum allowable hourly change in a unit’s output, minimum and maximum output when online, minimum up and down times when a unit is started of stopped, etc. Although most unit costs and constraints can be parameterized in this manner, some SO commitment models are too rigid to fully capture unit characteristics. Combined-cycle gas turbines, for instance, have “sawtooth”-shaped cost functions because they can switch between different operating modes as their output changes. Cascaded hydroelectric systems have constraints linking the reservoirs within the watershed, which some SOs cannot fully capture in their standardized constraint parameters.22 Given these inputs, the SO then determines a least-cost commitment and dispatch of the units to serve load while satisfying all operating, security, and ancillary service constraints. The SO will typically settle the market based on a uniform marginal energy price for each hour in the planning horizon at each location in the transmission network. These hourly locational marginal prices (LMPs) are paid to each unit that is dispatched to generate

EBL

19

In some markets, the SO uses its own load forecasts instead of soliciting bids from consumers in order to ensure a sufﬁcient set of units is committed. Many markets which do commit units on the basis of submitted demand bids have a separate reliability unit commitment process into which the SO substitutes its own load forecast and solves to ﬁnd if additional units (beyond those selected in the initial unit commitment) must be started to ensure its load forecast can be satisﬁed. 20 Some centrally committed SOs give generators the option to self-commit their units and submit self-schedules and one-part energy offers as opposed to having the SO make commitment decisions for them. Most generators in such markets have tended to use the centralized commitment with multipart offers, instead. One possible reason for this behavior is that a generator which self-schedules a unit must recover all of its variable generating and ﬁxed start-up and no-load costs through energy payments and will have to roll these ﬁxed costs into its one-part energy offers. Units which are centrally committed by the SO, by contrast, are given supplemental “make-whole” payments if energy payments do not cover all of their costs. This different treatment of self-scheduled and centrally committed units can serve to make a self-committed unit less attractive in the SO’s dispatch, since the energy cost of the unit will seem higher. Make-whole payments are described further later in this section. 21 SOs which include virtual bids in computing the unit commitment will often remove these bids in computing the reliability unit commitment, since virtual bids do not generally require actual physical delivery of energy. This is to ensure that there are sufﬁcient units available when the virtual bids are netted out in real-time. 22 Streiffert et al. (2005) further note that because the SO’s commitment problem does not decouple when the load balance and ancillary service constraints are relaxed, unit commitment problems solved by an LR algorithm cannot easily model combined-cycle gas turbines or cascaded hydroelectric systems. Instead, they must often rely on other approximation algorithms.

256

Competitive Electricity Markets

in that period. Because most units incur ﬁxed startup and no-load costs, these linear energy-only payments can be conﬁscatory – meaning that inframarginal energy rents may not ensure that a generator recovers its full costs. Since such conﬁscation could incent generators to withhold themselves from the commitment process or to misstate costs or operating constraint parameters in their offers, potentially affecting the efﬁciency of the resulting commitment, centrally committed markets typically include a supplemental “make-whole” provision. This make-whole provision pays any revenue shortfall over the course of the planning horizon to ensure committed units recover their stated costs. Units which receive energy payments in excess of their ﬁxed and generating costs do not receive any supplemental make-whole payments. Although the make-whole provision ensures units recover their stated costs over the course of the day, a unit may still be run at a loss in a single hour. This could occur, e.g., if a unit is kept online and generating at its minimum load (for instance, because of a minimum up-time constraint), in which case it would not set the LMP and could be running at a marginal loss.23 As such, units which do not lose money over the course of the day may be forced to cross-subsidize between individual hours. For ease of analysis and discussion the unit commitment formulation used in the simulations is a simpliﬁcation of a commercial SO model. It includes stepped marginal costs and ﬁxed (not time-dependent) start-up and no-load costs for each unit. Demand is priceinelastic; there are no network ﬂow constraints, virtual bids, self-schedules, or ancillary service requirements. Appendix 6A discusses in further detail the exact formulation studied. The computations assume the centrally committed market settles with a uniform hourly energy price – speciﬁcally the dual variable associated with the hourly load balance constraints.24 As noted before, because linear energy-only payments can be conﬁscatory, the analysis further assumes that the SO includes a make-whole provision, which pays each unit the difference between its total costs incurred in the commitment and dispatch (calculated on the basis of costs stated in its offer) and the total energy payments received over the course of 24 hours, if that difference is positive. These payments ensure that total net proﬁts (on the basis of stated costs) are always non-negative.

EBL

6.3.2. Self-committed model In a self-committed market, generators are left to make commitment decisions individually, as opposed to the SO making binding commitment decisions for them. The exact format and means of energy trade in self-committed markets varies, however. The current 23

A unit can also run at a marginal loss if it has any binding intertemporal (e.g., ramping) constraints and is the highest marginal cost unit, in which case it may not set the LMP. However, the LMPs used by the SO are typically computed from a separate optimal power ﬂow (OPF) problem, which determines the optimal dispatch of the units given the ﬁxed commitments from the unit commitment. Because these OPFs are solved separately for each hour, they do not include intertemporal constraints; thus a unit will only run at a marginal loss if it is held at minimum load. 24 As noted before, some SOs use LMPs based on the load balance constraint dual variables from the OPF, which can be different from the prices from the unit commitment load balance constraint dual variables. When there are no network ﬂow constraints, the OPF LMPs will simply be set by the highest marginal cost unit, which is not running at minimum load. The unit commitment prices, on the other hand, will “smooth out” prices when high-cost units have binding intertemporal ramping constraints, and can in general be different. Because the two sets of prices are nearly identical in these simulations, the analysis uses only the unit commitment prices.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

257

Texas market, for instance, is based almost exclusively on bilateral transactions between generators and consumers. The original California market operated as a day-ahead energyonly market into which generators would offer their generation and are given energy payments only for their actual generation. The California market made dispatch decisions solely on the basis of the generators’ energy offers by aggregating them into a supply curve and intersecting them with demand, without any regard for operating constraints or start-up or no-load costs. These constraints and ﬁxed costs were meant to be internalized by generators, which would be compensated for energy generated only.25 The model analyzed assumes that energy is traded through an energy-only market, such as the original California market design. To compute a competitive benchmark, the market is modeled as a competitive auction in which the auctioneer26 announces a set of hourly energy prices, and price-taking generators individually determine their hourly commitments and output level to maximize proﬁts and submit offers to the auctioneer indicating how many MWh they are willing to supply in each hour. The auctioneer then iteratively adjusts the hourly energy prices until it ﬁnds a set of prices which incent sufﬁcient generation to serve the load. This iterative price-updating process is meant to mimic Wilson’s (1997) proposal for a self-committed market with two important differences. One is that loads are ﬁxed in each hour as opposed to being price-elastic. Thus, the market is assumed not to accept demand bids but rather solicit sufﬁcient generation at any price to serve a ﬁxed hourly load. The other is that under Wilson’s proposal, generators are assumed to submit offers consisting of quantity/price pairs. Because the model analyzed in this section assumes generators to behave competitively, generators are modeled as price-takers, which take the auction prices as ﬁxed and decide their commitments and generation offers to maximize proﬁts individually, as opposed to strategically adjusting their energy offers to raise energy prices. Although the model assumes that energy is traded through a centralized energy market, it can be thought of as solving for a competitive equilibrium of direct bilateral trade between generators and consumers a la a Walrasian auction model. The model further assumes that the auctioneer starts with a set of prices which incent sufﬁcient generation to serve the load, and iteratively adjust prices until ﬁnding a set of supporting minimal prices – which is a set of prices such that generators offer sufﬁcient energy to serve the load, but would no longer do so if any of the energy prices were reduced. As the energy prices are dropped, higher-cost units will no longer ﬁnd it proﬁtable to commit themselves and the total quantity offered for generation will be driven toward the system load. One difﬁculty with ﬁnding a set of supporting minimal prices is that the binary nature of the generators’ commitment decisions means that a set of supporting minimal prices will generally not be market-clearing, meaning generators will offer more total generation than there is load to serve, yet reducing any of the energy prices will cause a unit to de-commit itself leaving insufﬁcient energy to serve the load.27 One solution is to assume that the auction uses some type of rationing rule to determine how the load is divided amongst generators willing to commit themselves. Our model assumes, instead, that because highercost generators drop out of the commitment as prices are iteratively reduced, if there is

EBL

25

Generators could also offer capacity into separate ancillary service markets in which they could be paid for committed reserve capacity. 26 The auction can be thought of as being operated by the SO, or it can be a separate outside market. 27 This stems from the fact that if the SO unit commitment problem is solved using an LR algorithm, the dual solution will have a non-zero duality gap.

258

Competitive Electricity Markets

excess generation offered and multiple units are competing for the same load, the one with the lowest average cost over the course of the day will prevail. In modeling generators’ proﬁt-maximizing behavior, they are assumed to perfectly rationally expect the behavior of other generators and take into account the “winner determination assumption” in making their own commitment decisions. This is to preclude the possibility that a unit may commit itself in expectation of being dispatched but ﬁnds that it does not, resulting in a net proﬁt loss. This assumption is enforced algorithmically and described along with the formulation and solution algorithm in further detail in Appendix 6A. Finally, the model assumes that each generator acts independently in making its commitment decisions, as opposed to making commitment decisions for portfolios of generators being owned by generating ﬁrms. This assumption is made because the dataset used does not have unit ownership information, although the technique and results would translate to a setting with generator asset portfolios. 6.3.3. Market simulations The simulations are based on actual market data from an ISO New England (ISONE) unit commitment problem in February 2005, consisting of 276 dispatchable units. This dataset is used because of availability of the data and is meant to be an illustrative example of the

EBL

Fig. 6.1. ISO New England system map.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

259

relative efﬁciency losses and settlement costs between the two market designs. Fig. 6.1 shows the ISONE system map and the nine major pricing zones. The ISONE system covers approximately 6.5 million retails customers, includes more than 350 generators with 31 000 MW of installed capacity, an all-time peak load of 28 127 MWh, and $11 billion of annual energy trade. The operating costs and constraint parameters used in the market simulations are those that were submitted by generators to ISONE. The competitive benchmark assumption takes these generator-offered parameters as reﬂecting actual costs and unit operating constraints – thereby assuming away any incentive compatibility issues. The computation of the central commitment assumes generators will offer these actual cost and constraint parameters to the SO for use in its commitment problem, as opposed to strategically misstating them to increase proﬁts. The computation of the self-commitment assumes that generators behave as price-takers and maximize proﬁts with the same cost and constraint parameters. Table 6.2 compares the total settlements paid to generators, commitment costs, and proﬁts of the generators in the simulations of the two market designs. Although the centrally committed market is assumed to include a make-whole provision, the dataset is such that each generator receives sufﬁcient inframarginal rents to recover all its costs and no supplemental payments are required. Nonetheless, Fig. 6.2 shows that the set of supporting minimal prices found in the self-committed market far exceed the energy prices paid in the central unit commitment. Indeed, a critical assumption underlying a centrally committed market is that the SO can force cross subsidies of “losing” hours by proﬁts from other hours and has means of preventing generators from making adjustments to their assigned schedules. Fig. 6.3 shows the resulting load imbalances which would occur if generators could individually adjust their outputs to maximize proﬁts against the hourly energy prices – known as uninstructed deviations. Due to the potential for such deviations, SOs penalize such deviations in generation by requiring generators to buy or sell back their insufﬁcient or excess generation at the LMP, thereby removing any incentive for such deviations. This enforcement mechanism can be problematic, however, in multiple-settlement systems in which the SO computes different sets of LMPs at different intervals in real-time. Because there can be differences between the prices at which an uninstructed deviation is paid and penalized, a generator may be inclined to change its output if these price differences are predictable. Alternatively, O’Neill et al. (2005) propose a two-part nonlinear tariff, which prevents such deviations. Their proposed scheme pays the same uniform LMPs28

EBL

Table 6.2. Cost and proﬁt comparison of centrally and self-committed market designs Market design

Energy payments

Make-whole payments

Total settlements

Commitment costs

Total unit proﬁts

Central

$16 075 121

$0.00

$16 075 121

$5 758 201

$10 316 920

Self

$25 060 666

$25 060 666

$6 003 274

$19 057 392

55.90%

4.16%

84.72%

%-difference

28

Importantly, these LMPs must be the prices from the unit commitment problem, not from the associated OPFs. Moreover, the pricing scheme enforces the unit commitment dispatch, not that from the OPF. If the SO dispatches according to the OPF, then those OPF constraints must be incorporated into the unit commitment formulation.

Competitive Electricity Markets

260

15000

160

Load Central commitment energy price Self commitment energy price Self commitment marginal average cost

140

14000

Load (MWh)

100 12000 80 11000 60 10000

40

9000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

20

Hour Fig. 6.2. Central- and self-commitment prices of ISONE market simulations.

EBL 80 Load Imbalance

60 40

MWh

20 0 −20 −40 −60 −80

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Hour Fig. 6.3. Load imbalance from uninstructed deviation in centrally committed market.

Price ($/MWh)

120 13000

Cost of Anarchy in Self-Commitment-Based Electricity Markets

261

and a set of discriminatory payments on each of the commitment variables.29 O’Neill et al. demonstrate that such a pricing scheme will ensure that proﬁt-maximizing generators will follow the same centrally determined commitment and schedule if they could adjust their outputs, without the need for any additional penalty or enforcement mechanism.30 Inclusion of such a provision in our simulation would increase total settlement costs of the centrally committed market by approximately $219 017, which is a relatively small sum compared to the total simulated settlement costs of the market of approximately $16 million. More importantly, the simulation demonstrates that a self-committed market requires higher energy prices than a centrally committed one. Because the model assumes that demand is ﬁxed and inelastic, these higher prices are simply a wealth transfer from consumers to generators, without any efﬁciency losses. In a more realistic setting with demand response,31 the higher prices could result in allocative distortions. Table 6.2. also shows that a self-committed market will generally suffer productive efﬁciency losses, as demonstrated by the more than 4% increase in total commitment and dispatch costs. These efﬁciency losses are not the result of units committing themselves under a self-committed market when they would not be committed under the central commitment. Rather, these losses stem from the fact that a central commitment gives the most efﬁcient coordination of generator dispatches, which are lost when generators dispatch themselves independently. Of the 276 units, 108 are committed in at least 1 hour under the central unit commitment solution. Of these 108, 73 follow the same commitment and dispatch schedule under the self-committed market as under the central unit commitment, with some shufﬂing of generation amongst the remaining 35. 6.4. Conclusions

EBL

This chapter has revisited one of the key issues surrounding the proper role of the SO in the design of competitive electricity markets. An SO with broad economic authority can, in theory, determine the most efﬁcient commitment to meet forecasted demand. However, centrally committed markets, are not strategy-proof and are prone to incentive compatibility issues, meaning that generators can proﬁtably manipulate their offers to increase proﬁts. This has both been shown through simple examples and was one criticism of the original Electricity Pool in the British market. Proponents claim that a decentralized energy-only market in which generators individually determine their commitments can reduce the incentive issues of a central unit commitment while minimizing efﬁciency losses. The simulation of a competitive benchmark conducted in this chapter provides a bound on the productive efﬁciency losses from a self- as opposed to centrally committed market design. While these losses were relatively small, around 4.25% in the case examined here, this would nonetheless represent a signiﬁcant welfare loss in absolute terms considering that ISO markets typically trade energy worth billions of dollars on an annual basis. The efﬁciency loss in the market simulations would amount to an annual loss of nearly 29

These commitment variable payments are found by solving the unit commitment problem to optimality, adding constraints to ﬁx the values of the commitment variables, solving the linear programming relaxation of the unit commitment problem, and using the dual variables on the constraints ﬁxing the commitment variables. 30 Their results apply much more broadly to any market with integer variables or other nonconvexities. 31 See Chapter 8 for a discussion of demand response.

Competitive Electricity Markets

262

$90 million for the ISONE system, if the results are typical of most days.32 Moreover, the total efﬁciency losses would be signiﬁcantly higher in the presence of demand response since the energy prices under self-commitment will generally be higher than under central commitment. These higher prices could cause self-committed markets to be more prone to allocative efﬁciency losses in the presence of demand response, which may be an important consideration as these programs are slowly becoming more prevalent. Although these simulations are based on a dataset from a single SO for a speciﬁc period of time, the results are nonetheless instructive for estimating the relative size of the efﬁciency losses between the two market designs. The issue of unit commitment is an important and often overlooked one. Many authors have been quick to advocate one market design or another, all the while suppressing these important technical realities of power systems. The discussion and analysis of this chapter and the issues raised would be relevant to any market that is evaluating the options of centralized versus decentralized design of day-ahead markets. In some sense, the options available to policy makers and market engineers is between two imperfect systems, since centralized markets will be fraught with incentive problems and decentralized markets with coordination losses. 6A. Appendix This section describes the speciﬁc formulations and algorithms used in modeling of the centrally and self-committed market. 6A.1. Central Commitment Model

EBL

To describe the formulation, the following notation is used: Problem Parameters • • • • • • • • • • • • •

I: generator index set T: number of planning periods B: number of steps in generators’ marginal cost SUi : stated start-up cost of unit i Ni : stated no-load cost of unit i MCbi : stated marginal generating cost of step b of unit i’s marginal cost curve bi : stated maximum generating capacity of step b of unit i’s marginal cost curve − Kit : stated minimum generating capacity of unit i in period t + Kit : stated maximum generating capacity of unit i in period t Ri : stated maximum ramp rate of unit i ni : stated minimum up time of unit i fi : stated minimum down time of unit i Di : load forecast in period t

Decision Variables • • • • 32

b qit : generation provided from step b of unit i in period t uit : binary variable indicating if unit i is up in period t sit : binary variable indicating if unit i is started in period t and hit : binary variable indicating if unit i is stopped in period t.

This may likely be a lower-bound on the annual losses, since more energy would presumably be traded during summer peak periods.

Cost of Anarchy in Self-Commitment-Based Electricity Markets

263

+ − , Kit , Ri, ni , and fi , are The generator cost and constraint parameters, SUi , Ni , MCbi ,bi , Kit assumed to be given by each generator to the SO in the generator’s offers. Given these offers, and the load forecast, the SO’s unit commitment problem is then formulated as follows: b MCbi qit + Ni uit + SUi sit minimize total costs min qush

s.t.

it

ib

b

b qit

= lt

∀t

load balance

∀i t

unit generating capacity

∀i t b

unit segment capacity

∀i t

ramping limit

si ≤ uit

∀i t

minimum uptime

hi ≤ 1 − uit

∀i t

minimum downtime

∀i t

startup transition

− uit ≤ Kit

b

+ b qit ≤ Kit uit

b ≤ bi qit

−Ri ≤

t =t−ni +1 t =t−fi +1

b

b b qit − qit−1 ≤ Ri

sit ≥ uit − uit=1 hit ≥ uit−1 − uit uit sit hit ∈ 0 1 b ≥0 qit

EBL

∀i t

shutdown transition

∀i t

integrality

∀i t b

non - negativity

6A.2. Self-commitment model The self-commitment model assumes that the market operates as an iterative energy-only auction. The auctioneer announces a set of hourly energy prices, given by pt . Generators then individually determine their proﬁt-maximizing commitments and dispatches and submit to the auctioneer a set of offers indicating how many MWh of energy they are willing to generate in each of the hours. The auctioneer then iteratively adjusts the prices until reaching a set of supporting minimal prices. The “winner determination rule,” which states that if two or more generators are contending for the same dispatch the one with the lowest average cost will prevail, is enforced algorithmically. The model assumes that generators, in making their commitment and dispatch decision, will be perfectly rational in predicting other generators’ behavior and the winner determination rule outcome. This is to preclude the possibility that a generator commits itself thinking that it will be dispatched, but ﬁnding that it is not because of the commitment decision of another unit with a lower average cost. This is enforced in each generator’s proﬁt-maximization problem by restricting each generator to produce no more than the available load in each hour, where the available load in each hour is the total load in that hour less the dispatch of lower average cost units. As units are accepted for dispatch in order of average cost, the available load is updated, and the proﬁt-maximization problem of the remaining generators are iteratively resolved.

Competitive Electricity Markets

264

Using the same notation as from Appendix 6A.1, deﬁne pt to be the hourly energy prices announced by the auctioneer and alt to be the available load in each hour. Generator i’s proﬁt-maximization problem is then formulated as follows: b − Ni uit − SUi sit maximize proﬁts pt − MCbi qit Pt max qush t

s.t.

b

∀t

available balance

∀t

unit generating capacity

∀t b

unit segment capacity

∀t

ramping limit

sit ≤ uit

∀t

minimum uptime

hit ≤ 1 − uit

∀t

minimum downtime

sit ≥ uit − uit=1

∀t

startup transition

hit ≥ uit−1 − uit

∀t

shutdown transition

∀t

integrality

∀t b

non-negativity

it

b qit ≤ alt

− uit ≤ Kit

b

+ b qit ≤ Kit uit

b ≤ bi qit

−Ri ≤

t =t−ni +1 t =t−fi +1

b

b b qit − qit−1 ≤ Ri

uit sit hit ∈ 0 1

EBL

b ≥0 qit

Given a set of energy prices, pt , the resulting dispatch is found using the following algorithm: initialize alt ← Dt ∀ t = 1 T # initialize available load in each hour to be total load in that hour dispatchi ← 0 ∀i ∈ I # initialize each generator’s status to not being dispatched repeat ∀i s.t. dispatchi = 0 solve Pi # solve each undispatched unit’s profit-maximization problem b b MCi qit + Ni ui + SUi sit t b ∀i s.t. dispatchi = 0 let avgcosti ← b qit tb

Cost of Anarchy in Self-Commitment-Based Electricity Markets

265

# calculate each undispatched unit’s average cost b Let j ← arg minavgcosti dispatchi = 0 qit > 0 i

tb

# find undispatched unit with positive generation offer and lowest average cost. Denote this unit as unit j Fix commitment and dispatch of unit j Let dispatchj ← 1 # set generator j’s status to being dispatched b ∀t = 1 T Let alt ← alt − qjt b

# update available load in each hour to be previous available load, less generation offer of unit j until alt = 0 ∀t = 1 T or dispatchi = 1∀i ∈ I # repeat until available load in each hour is zero or all units have been dispatched

The algorithm is designed to ﬁnd a set of proﬁt-maximizing commitments and dispatches, while also ensuring that generators properly take into account how much load will be available for them to supply in each hour given the behavior of other generators and the winner determination assumption. It does so by iteratively accepting generators one at a time in order of their average costs, updating the load available in each hour based on the dispatch of the accepted units and resolving the proﬁt-maximization problems of the uncommitted units. The process continues until all the load is served, in which case the energy prices are called a feasible set of supporting prices, or until no other generators wish to commit themselves, in which case the energy prices do not incent sufﬁcient commitment to serve the load. To ﬁnd a set of supporting minimal prices, the auction begins with a set of supporting prices that will induce oversupply. The prices are then iteratively updated to be pt ← pt + 1 − t , where t is the highest average cost of the units dispatched to generate in period t and ∈ 0 1is a step-size parameter. In doing so, the auction aims to drive the energy price in each hour toward the marginal average cost. Once the prices cannot be feasibly updated in this manner, the auction decreases the energy price in each period individually so as to minimize the total settlement costs paid by the SO.

EBL

References Adib, P. and Zarnikau, J. (2006). Texas: The most robust competitive market in North America. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Amundsen, E.S., Bergman, L., and Von Der Fehr, N.-H.M. (2006). The Nordic electricity market: Robust by design? In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Bertram, G. (2006). Restructuring the New Zealand electricity sector 1984–2005. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier.

266

Competitive Electricity Markets

Bowring, J. (2006). The PJM market. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. California ISO (2007). Business Practice Manual for Market Operations. http://www.caiso.com/ Dyner, I., Arango, S., and Larsen, E.R. (2006). Understanding the Argentinean and Colombian electricity markets. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Guan, X., Zhai, Q., and Papalexopoulos, A. (2003). Optimization based methods for unit commitment: Lagrangian relaxation versus general mixed integer programming. In Proceedings of IEEE Power Engineering Society General Meeting. Hobbs, B.F., Rothkopf, M.H., O’Neill, R.P., and Chao, H-P. (eds) (2001). The Next Generation of Electric Power Unit Commitment Models. Norwell, MA: Kluwer. Hogan, W. (1994). An efﬁcient bilateral market needs a pool. Testimony before the California Public Utilities Commission. Hogan, W. (1995). Coordination for competition in an electricity market. Working Paper. Hortaçsu, A. and Puller S.L. (2005). Testing Models of Strategic Bidding in Auctions: A case study of the Texas Electricity Spot Market. UCEI CSEM Working Paper 125. Hunt, S. (2002). Making Competition Work in Electricity. New York: John Wiley & Sons. ISO New England (2007). Manual for Market Operations. http://www.iso-ne.com/ Johnson, R.B., Oren, S.S., and Svoboda, A.J. (1997). Equity and efﬁciency of unit commitment in competitive electricity markets. Uti. Pol., 6, 9–19. Joskow, P.L. and Schmalensee, R. (1983). Markets for Power: An Analysis of Electric Utility Deregulation. Cambridge, MA: The MIT Press. Mas-Colell, A., Whinston, M., and Green, J. (1995). Microeconomic Theory. New York: Oxford University Press. Midwest ISO (2005). Business practice manual for energy markets. Moran, A. (2006). The electricity industry in Australia: Problems along the way to a national electricity market. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Muckstadt, J.A. and Koenig, S.A. (1977). An application of Lagrangian Relaxation to scheduling in power-generation systems, Op. Res., 25, 387–403. New York ISO (2001). Manual 11: Day-ahead scheduling manual. Newbery, D.M. (2006). Electricity liberalization in Britain and the evolution of market design. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. In Press O’Neill, R.P., Sotkiewicz, P.M., Hobbs, B.F., et al. (2005). Efﬁcient market-clearing prices in markets with nonconvexities. Eur. J. Op. Res., 164, 269–85. OFFER (1998). Review of electricity trading arrangements. Ofﬁce of Electricity Regulation. Oren, S.S. and Ross, A.M. (2005). Can we prevent the gaming of ramp constraints? Dec. Supp. Sys., 40, 461–71. Raineri, R. (2006). Chile: Where it all started. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Amsterdam: Elsevier. Ruff, L.E. (1994). Stop wheeling and start dealing: resolving the transmission dilemma, Elec. J., 7, 24–43. Sioshansi, R., O’Neill, R.P., and Oren, S.S. (2007). Economics consequences of alternative solutions methods for centralized unit commitment in day-ahead electricity markets. Forthcoming in IEEE Transactions on Power Systems. Streiffert, D., Philbrick, R., and Ott, A. (2005). A mixed integer programming solution for market clearing and reliability analysis. In Proceedings of IEEE Power Engineering Society General Meeting, San Francisco, CA. Sweeney, J.L. (2006). California electricity restructuring, rhe crisis, and its aftermath. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds.). Amsterdam: Elsevier. Wilson, R. (1997). Activity rules for a power exchange. In Proceedings of the 2nd Annual POWER Conference of the University of California Energy Institute, Berkeley, California. Wilson, R. (2002). Architecture of power markets. Econometrica, 70, 1299–340. Wolsey, L.A. (1998). Integer Programming. New York: Wiley-Interscience.

EBL

Chapter 7 Market Power and Market Monitoring PARVIZ ADIB1 , AND DAVID HURLBUT2 Automated Power Exchange, Santa Clara, California, USA; 2 National Renewable Energy Laboratory in Golden, Colorado, USA

1

Summary Competitive electricity markets are prone to abuse of market power by dominant generators or ﬁrms with strategic assets under a number of circumstances. Experience suggests that when these conditions are present and a vigilant watchdog is absent, ﬁrms may be tempted to exercise their market power to manipulate prices in a number of ways, harming consumers and other market participants in the process. This chapter provides an overview of how market power can be exercised and how effective market monitoring can prevent abusing practices.

EBL

7.1. Introduction Competition is good public policy as it encourages efﬁciency and price transparency. While proper market design is important, the success of competitive electricity markets depends ﬁrst and foremost on structural conditions that do not distort market outcomes. One of the important conditions is that structural market power must be eliminated or effectively mitigated.1 Other structural conditions such as demand response are related to market power. The ability of customers to respond to high prices by reducing consumption can greatly limit suppliers’ ability to inﬂuence prices. Unfortunately, in most restructured electricity markets today, some degree of market power continues to exist. Demand response is still in its infancy, which means that customers have little ability to defend themselves when a dominant supplier engages in the exercise of market power.2 1 Structural market power refers to high concentration of ownership. Some wonder why market power abuse seems more prevalent in the United States than in Australia. One important difference between, for example, the Texas market and the Australian market is the high concentration of ownership in Texas, where TXU owns 22.4% of ERCOT-wide generation and 44.4% of generation in the transmission constrained North zone, versus the dispersion of ownership in Australia where the largest generation company has a 14% market share. See Moran and Skinner, Chapter 11, this volume. 2 See Chapter 8 by Jay Zarnikau in this volume for a detailed discussion of demand response in restructured electricity markets.

267

268

Competitive Electricity Markets

Until demand response matures, market monitors have a crucial role in detecting and reporting market power abuses. See Mark Armstrong and David Sappington (2006). Wholesale electricity markets have characteristics not normally encountered in antitrust law. Section 7.2 of this chapter elaborates on what market power means in the particular context of wholesale electricity markets. The section provides a review of the reasons why antitrust law is not always applicable to restructured electricity markets and why it cannot effectively address some of the market power problems encountered in those markets. The different types of market power and the different ways a monopolist can withhold production to affect market prices are described, and the tools available to market monitors to detect market power abuses and mitigate their impact are discussed. Section 7.3 provides a description of the role and function of market monitoring units (MMUs) that have been created in the United States and elsewhere. Section 7.4 provides conclusions. 7.2. Market Power in Electricity Markets Structural market power exists where a single player or a group of colluding players is large enough to control market outcomes. In wholesale power markets, a player is said to have market power if it is pivotal; which means that it is so large relative to the pool of other suppliers and to demand that, without it, supply would be physically less than demand and the market would not clear. There are no potential substitutes for electricity at least in the short run, resulting in what could be aptly termed the “Don Corleone Equilibrium”: the largest player can make the market an offer it can’t refuse. An important distinction is often made between having market power and abusing or exercising it. For example, a particularity of Texas law relating to the opening of a competitive electricity market is that it makes such a distinction between having and abusing market power. The Texas law deﬁnes market power abuse as “practices by persons possessing market power that are unreasonably discriminatory or tend to unreasonably restrict, impair, or reduce the level of competition.” Texas Public Utility Regulatory Act (2005): (Emphasis added.) This distinction means that abusively exercising market power is a practice that is illegal in Texas, whereas merely having market power is tolerated. In contrast, there are jurisdictions where the statutes address not only the practice but also the possession of market power. In such markets, effective structural remedies to address market power such as divestiture have been implemented. Thus, in California, New York, and New England, regulators have either required large vertically integrated utilities to divest their generation assets before the implementation of deregulated markets or provided strong incentives such as predicating stranded investment recovery upon asset sales. Divestiture of assets is the most efﬁcient method of mitigating the natural tendency of a monopoly company to use its market power to inﬂuence market prices and eliminate competitors. Whereas vertical divestiture is called for to separate the generation part of the former utility from the wires company and the retail branch, horizontal divestiture ensures the breakdown of large generation ﬂeets and their acquisition by multiple new owners.3

EBL

3 California’s three investor-owned utilities sold their generation assets to two or more buyers, which ensured better mitigation of their generation monopoly. In New England, however, regulators did not impose a limit on the maximum amount of generation that could be purchased by any single entity, and almost all of NEES non-nuclear generation (4000 MW) was acquired by a single buyer, PG&E, so that the potential for horizontal market power in that market remained strong.

Market Power and Market Monitoring

269

In electricity markets where one or more generation companies continue to have market power, the regulator’s and market monitor’s most challenging task is to protect customers from the high prices induced when market power is exercised. This section examines the speciﬁcities of horizontal market power when it is exercised in electricity markets and is organized into the following: • •

A discussion of the limitations of antitrust law when applied to electricity markets. A discussion of the distinction between high prices caused by a shortage of supply compared to the demand, referred to as scarcity prices, and excessive prices that result from the exercise of market power. • An examination of the different ways an entity with market power can withhold resources to control prices. 7.2.1. Antitrust law and electricity markets While antitrust case law contains a number of deﬁnitions of market power that have passed judicial muster, many have been tailored to the facts of the speciﬁc case being litigated. Antitrust and wholesale electric market oversight share the same principles: genuine competition is desirable, market power is not. These shared principles are rooted in fundamental market economics, in which a fair market environment engenders creative tension among economic actors. With respect to the speciﬁc problems associated with competitive electricity markets, however, antitrust offers few useful precedents. One expert in antitrust law observed that:

EBL

Experience over the last 15 years reveals that both regulation and antitrust may be hobbled in their ability to effectively deal with the competitive issues dealt out by restructuring. Antitrust may be limited because it cannot (by design) address certain problems; regulation, because it uses an inadequate set of policy tools.4 Generation owners usually advocate a verbatim adoption of the deﬁnition of market power used by the US Department of Justice (DOJ) and the Federal Trade Commission (FTC) in these agencies’ Merger Guidelines (DOJ and FTC 1997). Yet DOJ legal experts have said informally that they themselves do not apply the Merger Guidelines’ deﬁnition in such a way – as the title plainly states, they are guidelines, not laws or court rulings. They are designed to evaluate mergers, not market power issues that do not involve mergers. The DOJ document is therefore highly instructive, but by no means deﬁnitive outside the realm of mergers. The Merger Guidelines deﬁne market power as “the ability proﬁtably to maintain prices above competitive levels for a signiﬁcant period of time.”5 However, the element of time renders part of this deﬁnition less meaningful for the wholesale electricity market.6 The main question for antitrust is whether a merger, which is a discrete and easily observable change, causes a competitive market to become measurably less competitive. In this context, “signiﬁcant period of time” is meaningful, because time is easily divided into “before” and “after” the merger. On the other hand, there is no discrete dividing line between “before” and “after” in a newly restructured electricity market, because the market has no history before the day the 4

Moss (2005), February, p. 3. DOJ and FTC (1997), p. 2. 6 For a concise discussion of why the time element is problematic, see Stoft (2002), pp. 365–8. 5

270

Competitive Electricity Markets

market opens to competition. Market power may have been present from the beginning, but the tools for detecting it and measuring it need to evolve.7 Time also has a different meaning in markets in which customers have the ability to see prices when they make consumption decisions and to defer consumption accordingly, and where consumers and suppliers have the ability to store products for future use or sale. In contrast, a large majority of electricity customers are blind to price ﬂuctuations and cannot respond to them, are unable to defer or forego consumption at least in the short run, and cannot store electricity for future use. Hence relying on the Merger Guidelines to address market power in electricity markets would most assuredly cause an agency charged with market oversight to overlook the existence of market power. In short, while antitrust law may provide philosophical guidance, there is no basis for being bound to its precedents.8 Where market power exists, the challenge for a regulatory authority charged with protecting the public is to adopt a deﬁnition of market power that recognizes the complexities of electricity markets. It may be a highly generic deﬁnition taken from antitrust law that addresses the fundamental concerns, such as “the ability to control prices or exclude competition in a relevant market.”9 This deﬁnition of market power does not include the element of time found in the DOJ and the FTC deﬁnition mentioned above, nor does it require the market monitor to demonstrate that a company was able “proﬁtably to maintain prices above competitive levels.”10 A general deﬁnition, however, places the burden on market monitors and enforcement authorities to defend in court the tests and criteria they use to measure “the ability to control prices” if such tests are not included in the deﬁnition.

EBL

7.2.2. Scarcity prices versus high prices induced by market power In a well-functioning competitive market, scarcity puts pressure on prices, and the resulting high prices provide a signal for resource allocation decisions that bring about adjustments in demand and, over time, new investment in generation. It does not follow, however, that any high price is a scarcity price. High prices can occur from the exercise of market power in the absence of true scarcity. Unlike scarcity prices, prices that rise due solely to the willful effort of one or more dominant players do not send rational signals for resource allocation and for that reason can be harmful to the market. Marginal costs and long-term ﬁxed costs can vary widely with the type of generation. In a uniform pricing auction where generators tend to bid in their marginal costs, generators with the lowest marginal costs are selected ﬁrst and those with the highest marginal costs are selected last. In most auctions, the marginal unit sets the market-clearing price or MCP. The marginal unit is the most expensive unit selected, and the MCP is the price at which 7

Moss (2005), pp. 33–4. This standard was applied by the court to the Texas Commission’s enforcement rule, Subst. R. 25.503. TXU Generation Co. v. Public Utility Commission, 165 S.W.3d 821 (Tex. App. – Austin 2005, writ denied). 9 This deﬁnition is drawn from the seminal US Supreme Court decision in U.S. v. E.I. duPont de Nemours & Co., 351 U.S. 377, 76 S.Ct. 994 L.Ed.2d 1264 (1956). It was adopted by the Public Utility Commission of Texas (25 TAC Sec. 504), with the intent that speciﬁc indicators of market power and their validity be considered in the context of speciﬁc cases as they arise. 10 DOJ and FTC (1997), Horizontal Merger Guidelines. “For more background on how DOJ and FTC employ these concepts in an investigation, see DOJ, commentary on the horizontal merger guidelines,” March 2006. 8

Market Power and Market Monitoring

271

all suppliers are paid.11 Under this pricing mechanism, all generators except the marginal unit receive a price that is higher than their marginal cost. This difference is known as the inframarginal proﬁt. Since marginal costs cover only short-term production costs, additional revenues are needed to cover the generator’s long-term ﬁxed costs and ensure ongoing investments in generation. Thus, for the business to be viable, the inframarginal proﬁt must provide the revenues necessary to cover both the short-term and long-term costs of electricity production. Scarcity refers to a situation where the amount of generation capacity needed to serve the load and to maintain a minimum reserve margin approaches existing capacity, putting pressure on prices. High MCPs lead to high inframarginal proﬁts, which, in turn, lead to investment in additional resources. The new plants entering the market tend to have low marginal costs compared to the existing ﬂeet. Once the older, more expensive units that had been setting the MCP are displaced by the new efﬁcient units, the next-most-expensive unit in the stack will set the price, with the effect that over time prices will come down. As long as the new price setters have marginal costs sufﬁciently higher than the new units, however, the new generator will be proﬁtable and long-term sustainability will be achieved. The previous paragraphs describe how price signals are expected to work in a wellfunctioning competitive market. However, when prices are driven by the exercise of market power rather than by scarcity of supply, the price signal is corrupted. Lenders and investors may be uncertain about whether to build or ﬁnance new plants if they see an incongruity between high prices and seemingly abundant supply. Prices that are higher and more volatile than the fundamentals of supply and demand would suggest can also be a signiﬁcant risk factor for new retail suppliers in the electricity market. In addition, if consumers cannot respond to high prices by reducing their demand, excess wealth ﬂows from consumers to producers. Market power also makes it more difﬁcult for intermediaries such as power marketers to arbitrage risk. Consider, for example, the different risk values placed on long-term contracts by a typical generator and by a typical retailer. For the generator, a 3-year contract to provide power reduces risk because it assures some amount of future revenue. The retailer, on the other hand, may not know what its customer base will be 3 years into the future, and thus may prefer short-term supply arrangements. Consequently, there is a natural economic role for a power marketer to arbitrage risk between the two market players, to the advantage of all involved. For power marketers to effectively perform this function, they must have conﬁdence that prices change in response to legitimate market mechanisms and will continue to do so. If the prices paid by the retailer and received by the generator respond to the normal forces of supply and demand, power marketers can safely base their market decisions on informed predictions about future prices. If price ﬂuctuations are instead the result of market power abuse, the prices do not follow a predictable course, the associated risk is difﬁcult to arbitrage, and the market operates less efﬁciently. Unfortunately, there is no simple test to distinguish a legitimate $1000 scarcity price from a $1000 price resulting solely from the exercise of market power. Sometimes both scarcity and market power may be at work. The data for supply, demand, market share,

EBL

11

Uniform pricing, as compared to pay-as-bid pricing, is the dominant pricing mechanism used in most restructured wholesale electricity markets, particularly, in the United States. For a good discussion of these two pricing mechanisms, see Kahn et al. (2001).

272

Competitive Electricity Markets

and prices may suggest that market power is what drives prices, but the difﬁculty comes in proving that a speciﬁc action by a speciﬁc market entity constitutes an exercise of market power. To compound the problem, if a large supplier controls the market, it need not break any established market rules to exclude competition or manipulate prices. This is why structurally competitive markets – i.e., markets that exhibit ease of entry and where no supplier or group of suppliers can control prices – are generally better at disciplining market participants, protecting consumers and producing efﬁcient pricing, output, and investment than is reliance on behavioral rules. Absent the best solution of genuine competition, the regulators and MMUs who must protect consumers against unjust prices have a difﬁcult challenge.

7.2.3. The different types of market power Wholesale electricity market exhibits market power in three ways. Ideally, each type should be addressed differently when developing legislation prior to deregulation, when designing the market and developing the market rules, and when developing mitigation plans or penalties for anti-competitive behavior. This section provides a discussion of the distinction between pervasive market power, temporal market power, and local market power, and of the difﬁculties in concretely delimiting each type. 7.2.3.1. Pervasive market power If a supplier is pivotal all the time, that entity’s market power is called pervasive. This is the classic kind of market power that is described in economic textbooks and discussed in antitrust literature. Pervasive market power can be difﬁcult to address. Take, for example, a market in which auctions to purchase and sell energy are conducted every hour. The largest supplier may be pivotal half the hours, but only close to being pivotal the rest of the hours. How “close” is too close? Are half of all hours enough to be signiﬁcant? The second, third, and fourth largest suppliers may also be pivotal, but only for a handful of hours. Does that mean these middle-sized players have market power too? If the largest supplier is pivotal only because all the old, inefﬁcient, and expensive units have not been turned on, is that entity considered to have market power? These are some of the issues the market monitors face when attempting to determine whether large generation companies, in a given market, have been exercising market power. The Electric Reliability Council of Texas (ERCOT)12 provides an instructive case study of the challenges. In the fall of 2004, the largest generation owner in the ERCOT market stated its intention to begin pricing the energy it offered into the balancing energy market at well above marginal cost, so that it could recover both marginal cost and its own estimation of the long-term ﬁxed costs associated with its gas turbine units. The supplier was pivotal in the balancing energy market signiﬁcantly more often than even the second-largest supplier. A subsequent investigation by the Public Utility Commission of Texas’s (PUCT’s) market monitoring consultant concluded that the supplier’s pricing strategy was not consistent with competition and contributed to a signiﬁcant increase in balancing energy prices

EBL

12

The ERCOT accounts for about 85% of Texas. The Public Utility Commission of Texas (PUCT) has regulatory jurisdiction over ERCOT wholesale electricity market. In contrast, the FERC has jurisdiction over wholesale electricity markets in all other states within the United States.

Market Power and Market Monitoring

273

during the period studied.13 Nevertheless, two factors prevented investigators from stating that market power had been exercised: •

First, other suppliers had excess spinning capacity that they could have offered into the market but did not. Had all of that capacity been offered at its marginal cost, the investigators concluded, prices would have remained normal. • Second, at the time the PUCT had adopted no particular deﬁnition of market power. The investigators, acting on their own judgment, used the DOJ/FTC deﬁnition previously mentioned, which requires proving that the entity had the “ability proﬁtably to maintain prices above competitive levels for a signiﬁcant period of time.” The analysis could not demonstrate that the company proﬁted overall during the study period by the higher prices, even though it could show that the supplier’s strategy caused prices to be signiﬁcantly higher.14 No punitive action was taken at that time. However, a few months later in the summer of 2005, the company again started offering large amounts of balancing energy well in excess of marginal costs over a period of 4 months. A new investigation was launched. The supplier was found to be pivotal in about 84% of the price spike intervals, and it was again found that its pricing strategy was inconsistent with competition. This time, however, the investigator found that the company had proﬁted by close to $20 million over the 4 month period while raising prices by approximately $70 million in the balancing energy market, and concluded that the company had engaged in economic withholding and abused its market power.15 Thus, although the company engaged in the same pricing strategy in the two periods, based on the market power deﬁnition adopted, a determination of market power abuse could not be made conclusively in the ﬁrst investigation. This example illustrates the importance of establishing an operational deﬁnition of market power that is suited to wholesale electricity markets. Texas is not the only market where market monitors have been unable to demonstrate that market power has been exercised until recently. None of the MMUs that are members of the Energy Intermarket Surveillance Group (EISG)16 have reported any success in investigating and documenting market power abuses. As Leveque (2006) concluded recently, “The potential for market power in electricity and gas is high but extremely difﬁcult to determine.”17

EBL

13

Potomac Economics (2005). The reason was that the period of analysis was too short to make such a determination. During a small number of intervals the supplier had scheduled too little energy overall and ended up having to purchase balancing energy at the high price it had set. Over the short period of analysis, the magnitude of these few losses offset the majority of intervals during which the supplier received higher balancing energy revenues. 15 See Potomac Economics (2007). Investigation of the Wholesale Market Activities of TXU from June 1 to September 30, 2005. Independent Market Monitor, March. The report was further revised in September 2007 to reﬂect lower ﬁgures for proﬁt for $19 million and rising prices in balancing energy market by $57 million. The IMM report is available on the Texas PUC website: http://www.puc.state.tx.us or can be directly accessed at the Potomac Economics website at http://www.potomaceconomics.com/ercot/2005%20TXU%20Investigation%20(Final%20%20REDACTED).pdf. A Notice of Violation issued to TXU by the PUCT can be found on the website under Docket No. 34061. Interested readers may follow the progress of this case on the PUCT website under that docket number. 16 The EISG is a non-proﬁt, voluntary professional organization of market monitors from electricity markets around the world that was established in 2000. 17 See Levegue (2006), p. 31. 14

274

Competitive Electricity Markets

Market power often exists in degrees. Deciding how much is too much will necessarily involve subjective policy decisions that no amount of economic analysis alone can answer. What we can say without hesitation, however, is that competition is threatened when market power is so pervasive that certain suppliers can control prices even when reserve margins are high and there is no sign of scarcity. 7.2.3.2. Temporal market power Temporal market power is market power caused by a tightness of supply. In some contexts, the theory of contestable markets and the limitations on antitrust enforcement have supported the notion of allowing some level of market power to exist, as long as it is not accompanied by abusive behavior on the part of the monopolist.18 Under this theory, high prices, whether they are the result of scarcity or market power, will induce market entry that will correct the underlying deﬁciency and lower prices. In a practical sense, however, this low level of market power is barely distinguishable from scarcity. If market power does not exist or is of limited scale without scarcity, then the real issue is resource adequacy and not market power.19 The economic issue is not necessarily the resulting high prices, but rather the suppliers’ response to those prices. The types of market power abuse for which market monitors should be especially vigilant are activities by current suppliers that prevent new suppliers from entering the market. 7.2.3.3. Local market power Local market power is a common problem in electricity markets because of transmission bottlenecks, which are common in many markets today. Some states judged it prudent to direct former monopolies to divest their generation assets prior to opening the electricity market to competition. However, even after the former utilities divested their generation, these markets continued to experience local market power.20 Local market power arises from limitations on the transmission system. Line limits may make it impossible to import enough power into a load pocket to meet local demand, even if the system outside the load pocket has abundant generation resources. Under these conditions, a lone generator inside the load pocket may well be in a position to dictate its price with respect to the residual demand21 in the load pocket. Local demand cannot be served through imports entirely, and as a result, the local generator is in no danger of being replaced by low-cost competitors outside the load pocket. The “market” in which market power exists in this case is difﬁcult to deﬁne; it often has no geographic boundaries, but is rather delineated by a complex combination of power ﬂow relationships. Addressing local market power requires imposing mitigation safeguards. For example, it may be ruled that generators inside load pockets can only be paid their generic costs when conditions do not allow enough power to be imported into

EBL

18

See Baumol (1982). Also see Shepherd (1984) for a strong criticism of the theory of contestable markets. See Falk (2004) for an application of this argument to wholesale electricity markets. 19 For a more extensive discussion of scarcity and resource adequacy, see Adib, Schubert, and Oren, Chapter 9, this volume. 20 California is a good example. The state was successful in achieving both vertical and horizontal divestiture of its three largest utilities, yet the California electricity market was plagued by local market power problems caused by its severe transmission bottlenecks. 21 Residual demand is the difference between total demand for a given time period minus total offers by all other suppliers that should be served by the generator in question.

Market Power and Market Monitoring

275

the local area. The challenge comes from having to regulate prices only within the limits where local market power is likely to exist, without affecting other areas. Complicated modeling is often required to decide how to draw the lines and where to enforce the mitigation rules. Even then, any lines drawn are bound to be wrong some time or another. The question then, is whether to err on the side of too much mitigation or of too little: is it better to impose price limits when competition would otherwise be sufﬁcient, or to allow unrestrained pricing when in fact the market is not competitive? Generators are naturally predisposed to lobby for the latter, as it would provide occasional opportunity to reap a windfall by raising prices beyond competitive levels. However, erring on the side of too much mitigation has the least price impact on the market. Ideally, mitigation would be designed to make suppliers behave as though they faced competition. When that is possible, mitigation results in prices that are close to competitive prices and is least intrusive. A Texas example illustrates how market rules can attempt to mitigate local market power, but not always successfully. When local congestion occurs in ERCOT, a generator that is strategically located can submit a high premium bid22 knowing that the ISO has no other choice to resolve the congestion. To mitigate this problem, ERCOT protocols initially deﬁned a market solution as three unafﬁliated resources capable of solving a circumstance of local congestion, when no one bidder is essential to solving the congestion. Under this rule, when a market solution exists, each generator selected to resolve the congestion is paid the premium bid it submitted in the auction process, but when a market solution does not exist, each generator is paid at generic cost. This seemed at the time like a reasonable hands-off solution that would allow congestion constraints to be resolved by competitive forces when the condition for competition existed. However, market participants did not respond to competitive pressures as expected in situations where at least three non-afﬁliated competitors were offering their generation.23 Subsequently, the “market solution” provision was removed and it was decided that all generators called on to relieve local congestion would be paid at generic costs, with no exception.

EBL

7.2.4. Different ways to withhold resources Withholding of production is the most common method of exercising market power in a wholesale electricity market. Production may be withheld economically, which is referred to as economic withholding, by pricing it signiﬁcantly in excess of marginal cost. Or it can be physically withheld, which is referred to as physical withholding, by not offering it at all. All else being equal, however, and assuming no collusion, withholding is proﬁtable only if the supplier has market power, and only then does it become a concern. From the supplier’s point of view, market power greatly ampliﬁes the motive to withhold. Comparing inframarginal proﬁts demonstrates why. Inframarginal proﬁt is the difference between revenues and operating costs summed for all units in a supplier’s portfolio given a single MCP. Suppose a small supplier had a single 100-MW unit with 22

The premium bid was used by the ERCOT ISO to provide Balancing Energy Service when no Market Solution existed for resolving congestion. See 2003 version of the ERCOT Protocols §§ 2.1 and 6.1.11. 23 See Project No. 25937, PUC Investigation into Possible Market Manipulation of the ERCOT Market, “Settlement Agreement with Commission Staff,” 21 August 2003 and Memorandum to PUCT Commissioners, 21 August 2003.

Competitive Electricity Markets

276

a cost of $99/MWh, and that it was dispatched for all 100 MW. If the MCP were raised to $100/MWh, this supplier would proﬁt by only $100 ($1/MWh times 100 MWh). Now let’s assume a supplier with two units – one 50-MW unit with a cost of $99/MWh and a 50-MW unit with a cost of $20/MWh. This second supplier would proﬁt by $4050 if dispatched for that same 100 MW for an hour when the MCP is raised from $99 to $100. The larger and more diversiﬁed the portfolio is, the larger a supplier’s inframarginal proﬁt can be when prices are high. For a small supplier lacking the ability to control the MCP, maintaining such a high-cost unit would be risky. If prices were not high enough, the small supplier, if acting rationally, would eventually retire or sell the unit.24 On the other hand, a large supplier controlling a similar unit could, through withholding other units in its portfolio, make the price clear high enough to cover the unit’s operating cost, thereby signiﬁcantly increasing the proﬁts on all its other units. The large supplier, therefore, has the opposite incentive from the small supplier. Instead of retiring a unit that would be uneconomical in a competitive market, it would keep the expensive unit online as a potential price-setting tool. Figures 7.1 and 7.2 illustrate the two types of withholding, using a hypothetical market with 1200 MW of demand, 2000 MW of supply, and a large market participant that controls 50% of the supply. For ease of illustration, these examples assume that the large supplier’s cost curve is the same as that of the rest of the market. 7.2.4.1. Physical withholding In Fig. 7.1, the hypothetical large supplier signiﬁcantly increases its total revenues by taking three-fourths of its supply off the market. The overall supply shrinks, causing the market to clear at $325/MWh rather than $95/MWh. The large supplier sells all of the 250 MW it offered, but foregoes payment on 350 MW it would have sold had it not withheld. But because the price is more than three times higher, its total revenues are in fact greater.

EBL

Large supplier’s offer

Overall supply market

$750

$750

$500

$500

$250

$250

$0

Price increase due to withholding

$0 –

200

400 600 MW offered

800 1,000

Available MW offered Available MW not offered

–

500

1,000 1,500 MW offered

2,000

Supply, no physical withholding Supply with physical withholding MW required

Fig. 7.1. Physical withholding.

24 After enough retirements, competitive prices would eventually rise given tightness in the market through normal supply and demand, signaling new entry into the market.

Market Power and Market Monitoring Large supplier’s offer

277

Overall supply market

$750

$750

$500

$500 Price increase due to withholding

$250

$250

$0

$0 –

200

400 600 MW offered

800 1,000

Incremental cost Economic withholding offer

–

500

1,000 1,500 MW offered

2,000

Supply, no economic withholding Supply with economic withholding MW required

Fig. 7.2. Economic withholding.

(Its revenue under the withholding scenario is $81 250, rather than $57 000 under the no-withholding scenario. In addition, its costs would be lower under the withholding scenario, since it would be operating only 250 MW, rather than 600 MW.) Had a non-pivotal supplier with only 100 MW of supply attempted the same strategy, the effect on the clearing price would have been negligible. The supplier itself would have ended up worse off, having lost most of its potential revenue and most of the proﬁt it would have enjoyed had it priced its offer near its marginal cost. Hence the MMU does not have to be concerned with non-pivotal suppliers submitting offers at prices substantially above marginal costs.

EBL

7.2.4.2. Economic withholding In Fig. 7.2, the hypothetical large supplier doubles its revenues by pricing nearly all of its offers at $600/MWh. No other supplier has offered any amount priced higher than $500/MWh, so all 1000 MW offered by other suppliers are taken. Another 200 MW are required, however, and the large supplier is the only one who has it. The large supplier can name its price, and need not take its actual costs into consideration. Even though the large supplier sells a smaller quantity, it enjoys a much larger proﬁt on what it does sell due to the fact that its total revenue in this hypothetical scenario is $120 000 on total costs of $6500 for its least-expensive 200 MW. As in the example of physical withholding, if a small supplier were to increase its offer to $600/MWh, the impact on the MCP would be zero. The smaller supplier would simply price itself out of the market, giving up all the revenue and proﬁt it would have received had it offered at marginal cost. As these two examples illustrate, neither physical withholding nor economic withholding matters in the absence of market power. It is only when these activities are done by a large supplier with market power that the clearing price is signiﬁcantly affected. Consequently, the key element in a legal case involving market power abuse is not necessarily proving what the supplier did, but establishing that it has market power and that, therefore, what the supplier did led to an outcome that would not have occurred under competitive conditions.

Competitive Electricity Markets

278 7.2.5. Detecting market power

Market monitors screen for potential market power and possible market power abuse in two general ways: traditional methods that measure market concentration, and newer methods that test market outcomes. The traditional approaches, such as the Herﬁndahl– Hirschman Index (HHI) and the Delivered Price Test (DPT), are familiar to antitrust law. The outcome approach, however, has encountered rough waters in the courtroom despite its grounding in economic theory. This section brieﬂy reviews ﬁve methods of analysis that are widely used by market monitors to detect market power. The HHI and the DPT as well as the Pivotal Supplier Test (PST) indicate whether an entity has market power, but do not by themselves indicate whether the entity is abusing its market power. In contrast, the output gap test, the impact test, and scenario analysis provide information that indicates whether an entity with market power may actually be engaged in anticompetitive practices.

7.2.5.1. Concentration measures: HHI and DPT The HHI sums up the squared values of each supplier’s market share. For example, a market in which four suppliers each have 25% of the market would have an HHI of 2500 (25 squared, times four suppliers, where market share is measured as the number of percentage points). The DOJ uses the HHI to measure how market concentration might change as a result of a merger between two suppliers. Continuing with the above example, if two of the suppliers merged, the HHI would be 3750 (50 squared, plus 25 squared times the other two original suppliers). DOJ considers a market to be un-concentrated if the resulting HHI is less than 1000, and highly concentrated if the HHI is greater than 1800.25 The DPT as speciﬁed by federal statute is the HHI applied to a more explicitly delineated market.26 It evaluates the amount of capacity that can reasonably be expected to serve a market at a competitive MCP deﬁned to present 105% of that market’s prevailing price to attract additional potential supply, taking into account imports and native load obligations. When used to evaluate the market power effects of a merger or acquisition, the capacity identiﬁed by the DPT is then compared to the amount of capacity the entity will hold as a result of the merger or acquisition. The analyst is then able to assess if the amount of capacity held by the new merged entity will reach an unacceptable level as a percentage of the total capacity available to serve the market.

EBL

7.2.5.2. Pivotal supplier test As discussed earlier, a supplier is considered “pivotal” if it controls an amount of capacity that is so large relative to the demand and the pooled capacity of all other suppliers that the market cannot clear without it. Mathematically, a supplier is pivotal if the following equation holds: Qallsuppliers − Qpivotalsupplier < Demand 25 26

DOJ and FTC (1997), Horizontal Merger Guidelines, Section 1.51. See Title 18, Code of Federal Regulations, §33.3(c)(4).

Market Power and Market Monitoring

279

where Qallsuppliers represents the total amount of capacity offered to the market by all suppliers, and Qpivotalsupplier represents the amount of capacity offered by the entity being tested. If the total amount of capacity minus the capacity of the single supplier being tested is less than the demand, the single supplier is pivotal. This formula is the basis of the PST.27 A supplier that triggers the PST can unambiguously be said to have market power. It is important to note, however, that the supplier has market power at the time of the test and in the market where it is tested. Because the PST is calculated ex post, however, it might not be apparent that a supplier is pivotal until after the market clears and the opportunity to exercise market power has passed. Therefore, it is important, when conducting an analysis based on the PST, to assess how frequently and how predictably a large supplier might trigger the test. Each peak season will typically have a dozen or so hours when demand is so high that some moderate-sized suppliers may be found to be pivotal when applying the PST. In most cases, this is likely to be an example of temporal market power. These extraordinary periods are more indicative of thin reserve margins and true scarcity than of pervasive market power. A PST, therefore, should exclude these periods from the analysis up front or the analysis should give little weight to whether a supplier was pivotal during super-peak periods. A PST may be conducted on a system-wide basis by comparing real-time load against the generators that are online and capable of providing energy. It can also be applied to a speciﬁc market, as long as substitution with products outside the market is not reasonably possible. In the market operated by ERCOT, for example, market monitors have applied a PST to the balancing energy service (BES) market. Balancing energy is energy procured in real-time by the ISO to keep scheduled load and scheduled generation in balance. Generators offer into the BES market their excess spinning capacity (that which is not dedicated to serving load or other contractual obligation).

EBL

7.2.5.3. Output gap analysis The “output gap” measures the capacity of proﬁtable resources that are not offered into the market or are offered into the market at prices that are substantially above their marginal cost. For example, if a supplier were suspected of engaging in physical withholding, the market monitor would evaluate the amount of capacity this supplier had online but did not dispatch or offer into the market even though the marginal cost was signiﬁcantly less than the MCP. This amount of capacity constitutes the output gap. As discussed earlier, strategically keeping economic capacity off the market will shrink the supply of generation available to meet demand, forcing the ISO to procure higher-cost offers and thus increasing the MCP. The output gap analysis is useful as an initial screen for detecting possible withholding activities, but is not by itself conclusive evidence that strategic withholding has occurred. Many other factors can affect whether or not a unit is deployed. A full determination of market power abuse requires investigating and excluding each instance where a plausible, legitimate reason exists for not offering capacity, and a determination of how persistently the potential withholding occurred. 27 The results of a PST are sometimes presented as a residual supply index (RSI). The index can measure how frequently the market has a pivotal supplier, or how frequently any given supplier is pivotal.

280

Competitive Electricity Markets

7.2.5.4. Impact tests and scenario analysis Impact tests measure how market prices change as a result of behavior that fails a screen. The New York ISO, New England ISO, Midwest ISO, and PJM (Pennsylvania–Jersey– Maryland) ISO use some form of conduct and impact tests. A supplier fails the screen when it offers a generator’s capacity into the market at a very high price compared to an established reference price for that generator. When that happens, the bid price is administratively reset at the calculated reference price. However, these mitigation provisions are only implemented if the conduct has an impact on market prices, to ensure that seemingly high offers do not trigger price mitigation if they in fact have no impact on prices. Establishing reference prices for a particular generator requires an estimation of the generator’s marginal cost of production. The reference price is based primarily on fuel costs and incremental operating costs. When triggering conditions are met, i.e., when an offer is signiﬁcantly above its reference price, the security-constrained economic dispatch market pricing model can be re-run and the target generator’s offer is reset at its reference price. This allows the analyst to simulate the market outcome and price that would be in effect in the absence of anti-competitive behavior. The simulation indicates whether a supplier that behaved as though it had market power, i.e., priced well in excess of its marginal cost, actually affected the resulting market prices through its behavior. In practice, there are two ways price mitigation can be implemented. In one alternative, if the simulation provides evidence of market power abuse, bid prices can be reset ex ante, meaning that the bid price of a particular generator is administratively changed to the reference price before the results of the auction. Ex ante price mitigations are in effect in the New York and New England markets. Alternatively, if the results of the simulation show that the bidding behavior affects prices in excess of a threshold the prices obtained from the simulated run are used to clear the market. Although this approach may seem objective and data-driven, it in fact requires some subjective judgment. Most signiﬁcantly, there is little in the way of economic theory or legal precedent that can guide the selection of a threshold. Wherever it is set implicitly denotes an acceptable range of market power. For example, say a market participant failed an impact test with a $300 offer, but would not have failed had that same offer been $299. It would be difﬁcult for a market monitor to prove how a $1 price difference would constitute a material difference in the market participant’s ability to inﬂuence price outcomes. An example taken from the New England market illustrates another weakness of this approach in the calculation of a generator’s reference price. The New England ISO calculates a generator’s reference price by taking a rolling average of the bids submitted over time for this generator. The reasoning is that the bid price will reﬂect the generator’s marginal costs most of the time, except in certain situations when the supplier anticipates a shortage due to weather extremes or transmission or generation outages, or other similar event. As explained earlier, in the case when a generator’s bid price is signiﬁcantly above its reference price, the ISO either resets the bid price or replaces the MCP with a market price obtained by running a simulation model. In the spring of 2005, the New England ISO found that the reference prices calculated for certain non-competitive generation units that are frequently run out of economic merit to relieve transmission congestion were well above their marginal costs. These generators were never under competitive pressure to bid their marginal costs. The result was that their reference prices were well above their

EBL

Market Power and Market Monitoring

281

marginal costs. Once aware of the situation, the ISO in this case changed the rules and the way the reference price was calculated.28 7.3. Market monitoring Section 7.2 described the most important reason why electricity markets need close monitoring. The California crisis of 2000–0129 provided costly proof of how vulnerable consumers can be to market power abuse.30 Strategically located generation units resulted in local market power in areas where transmission constraints existed, and consumers could not respond effectively when prices soared. These structural problems were compounded by rules that prevented utilities from hedging against spot market volatility through longterm contracts, and by loopholes in rules that were intended to ensure reliability yet opened up unforeseen opportunities for gaming.31 32 The California crisis also illustrated how, when an MMU lacks coordination and close cooperation with the relevant enforcement agency, market power abuse can run unchecked, resulting in the failure of electric restructuring efforts.33 This section explores the status of market monitoring today and draws on the experience MMUs have accumulated with market power and market manipulations issues in California and elsewhere. The section is organized into: •

A discussion of the objectives of an MMU and of the ingredients necessary for meeting these objectives. • A review of the different ways MMUs have been organized in the United States and abroad, and of the extent to which organizational factors can affect the independence of an MMU and the resources it has available. • A review of how market monitoring has evolved since the beginning of restructured electricity markets, and of the adjustments that are being made to the structure and organization of MMUs to improve their effectiveness. • A summary of examples of conduct by market participants or the system operators that resulted in harm to the market.

EBL

7.3.1. Objectives of MMUs An MMU has three main objectives: identiﬁcation of market design ﬂaws, monitoring and reporting, and prevention. To meet these objectives, the MMU carries out the following functions. 28

For more details, see the following link: http://www.iso-ne.com/regulatory/ferc/orders/2005/ may/er05-767_5-6-05.doc. 29 See Borenstein (2002) and Borenstein et al. (2002) . For a description of gaming strategies, see Market Oversight Division(2002a). 30 Statistics collected by the California market monitor show that total electricity costs, which include energy and ancillary services costs at the wholesale level, increased from $5.5 billion in 1998 and $7.4 billion in 1999 to $27.6 billions in 2000. On a per unit basis, total costs went from $33 per MWh in 1998 and 1999 to $116 in 2000. 31 For a description of some of the gaming opportunities see Market Oversight Division (2002b). The two documents listed under this and the previous footnote were prepared by the authors of this chapter and other members of the Market Oversight Division, including Richard Greffe, Julie Gauldin, Tony Grasso, and Drs Eric Schubert and Sam Zhou. 32 See Sweeney (2006). 33 For an excellent discussion regarding market monitoring, see Wolak (2004).

282

Competitive Electricity Markets

Identiﬁcation of market design or operational ﬂaws: Poorly designed market rules and protocols can create unanticipated inefﬁciencies or gaming opportunities. One function of the MMU is to identify such inefﬁciencies and recommend changes to market rules to address them. The MMU also monitors ISO operations and evaluates the impact of the practices and procedures followed by the ISO on market outcomes and market efﬁciency. Monitoring and reporting on overall market performance: The MMU acts as a troubleshooter that identiﬁes trends that call for corrective action before a serious market failure occurs. It must detect market anomalies such as price excursions and determine if they are the result of market power manipulations or genuine scarcity. Prevention of harmful market activities: The market monitor can prevent market rule violations by identifying perverse incentives and recommending that they be changed. A market monitor that calls on market participants when they step outside legal limits can encourage compliance with market rules by simply letting it be known that someone is watching. Some of these harmful activities will be brieﬂy described later when real life examples of MMU investigations are discussed. It is not always easy to determine whether an event is the result of market manipulations or legitimate activities. The MMU will devote a lot of its resources to fact-ﬁnding activities and analyses of abnormal events before concluding that a potential violation has occurred. Communication with market participants is essential at this point, as what may appear to be a violation can often be explained by special circumstances or complex situations. Finally, the MMU may conclude that a certain behavior may have violated market rules, or may have been anti-competitive. If it does not have enforcement authority itself, it may, at that time, refer the case to the relevant enforcement agency. When that is the case, the MMU may act in a supporting role to help in the investigation phase, the negotiation phase, and the litigation phase as applicable.

EBL

7.3.2. Effective market monitoring To achieve the objectives described in the previous discussion, an effective MMU needs independence from stakeholders; appropriate skills, tools, and experience; adequate ﬁnancial resources; and access to both market data and individual market participants’ information. The independence of the market monitor will be discussed separately in the MMU organization discussion. The other ingredients to effective market monitoring are discussed below. The level of stafﬁng and budget of an MMU will vary depending on organizational factors. Some MMUs may rely on ISO staff for supporting activities related to information technology resources, human resources, and clerical support. The ERCOT Independent Market Monitor (IMM) has ﬁve full-time employees dedicated primarily to real-time market monitoring, market data analysis, investigations, report writing, or similar activities, but does not have enforcement authority and, therefore, does not have any legal staff. The IMM ﬁrst year budget is approximately $2 million. This level of stafﬁng and budgeting is considered at this time to be the minimum to conduct meaningful market monitoring in a market the size of ERCOT, but may have to be revised in future years if it is found that additional resources are necessary. Skills, tools, experience, and technical knowledge of system operations: Market monitoring requires extensive knowledge of economics, transmission operation, and system engineering. Tasks may also require sophisticated quantitative tools34 and computer 34

Quantitative tools are used to develop automated indices, screens, and reports (ISRs).

Market Power and Market Monitoring

283

models35 to replicate market operations and network topology and to allow scenario analysis. To be effective, the monitoring team needs expertise in economics, engineering, statistics, and operations research – preferably with several years of working experience. Frequently, market monitors will need to interpret and analyze massive amounts of market data. While some of the requisite knowledge may be obtained through education and work experience, most of the technical and operational experience can only be acquired through constant interaction with system operators. Financial resources: Many of the skills and experience needed for good market monitoring are also sought by market participants and are in short supply. The MMU therefore will have to offer high salaries to be competitive with the private sector. In addition, the MMU must budget for high-level consulting services, as the highly technical skills required for some monitoring and investigative activities will not always be available in-house. If the MMU is part of a governmental entity charged with market monitoring functions, budget restrictions may interfere with the ability to secure the required resources. While there are advantages to a market monitor located within a governmental enforcement agency, as will be discussed below, the inherent budgetary limitations are likely to severely impact the MMU’s effectiveness. Access to relevant market data and information: The market monitor can only function if it has access to all market and operational data. Therefore the market monitor has to have the authority to request all relevant information from the market participants whose behaviors are under investigation. Market participants will sometimes try to argue that they should not be required to provide sensitive market information for competitive reasons. An arrangement can and should always be made to accommodate the need for conﬁdentiality, and the MMU should be able to guaranty that the information provided will remain strictly conﬁdential, so this hurdle should not be difﬁcult to resolve. Timeliness is another important factor and the MMU must be able to place deadlines on its information requests, which can be negotiated so as not to impose an unreasonable burden on the market participant. Finally, it is essential that market participants’ data be archived for several years, as an investigation will often require looking into previous years’ activities. ISOs process thousands of pieces of market information all day long. Resource schedules and offers change hourly, there are new arrays of clearing prices every 5 to 15 minutes, regulation deployment every few seconds, and information regarding load forecast, generator outages, or tripped lines pours in all day long. Each time the market-clearing engine is run, the inputs are different and the outcome changes. This massive amount of data is where the evidence of market manipulations and market power abuse resides. The only practical way to examine this amount of data day to day is through computer-aided triage: automated tools and software routines that provide a user-friendly ﬁrst-cut analysis and ﬂag events that deviate signiﬁcantly from the norm. Such tools may include:

EBL

•

Market condition e-mail alerts (such as prices above a certain threshold level) and other frequently generated reports • Market statistics (price levels, congestion costs, outages, etc.) • Preliminary tests of market power and abnormal behavior (pivotal supplier, withholding, etc.) 35

Computer models may include statistical and econometric models as well as mathematical models, such as MCP engine, which allows MMUs to test speciﬁc behavioral changes by various market participants and evaluate how the market outcome might have been affected.

Competitive Electricity Markets

284

An MMU conducts day-to-day monitoring as well as long-term analyses, primarily examining the activities of market participants but also evaluating the ISO’s performance with respect to market operations, congestion management, and reliability. Table 7.1 provides a summary of these activities.36 7.3.3. Organization of MMUs The way an MMU is organized will often affect whether it can secure the necessary resources previously described, but also how independent it will be from market Table 7.1. Market monitoring activities Day-to-day • • • • • •

Real-time market monitoring through reliance on automated indices and screens Assessment of generation adequacy to meet forecast demand in various markets Review of abnormal events (weather, congestion, outages, etc.) Identiﬁcation of price abnormalities and conducting preliminary analysis Review of ex ante price mitigation and compliance Identiﬁcation of violations of market rules and procedures

Communication with market participants regarding abnormal outcomes and identifying market outcomes requiring further investigation • • • •

Daily reports preparation Databases and computer models maintenance Tracking of emerging market issues Dissemination of useful market information

EBL

Long-term • • • • • •

Identiﬁcation of market design ﬂaws and remedies Investigation of market power abuses and anti-competitive behavior Assessment of demand growth, generation and transmission adequacy, and system reliability Resolution of complaints and disputes regarding anti-competitive behavior Assessment of market competitiveness Preparation of regular reports (annual state-of-the-market reviews, monthly summaries, etc.) and reports on speciﬁc special issues and events

Monitoring of market operator* • • • • • ∗

Assessment of ISO compliance with market rules and operating guides and procedures, particularly those governing market operations and congestion management Detection of operational procedures and practices that create market inefﬁciencies or otherwise adversely impact market outcomes Determination of ISO’s short-term load forecast accuracy Evaluation of amount of ISO procured ancillary services Evaluation of ISO long-term planning process to identify demand growth and possible shortcomings in transmission system

Discussion by Adib (2006).

36

See also PUCT Substantive Rule §25.365 for a description of the market monitor’s role and responsibilities in Texas.

Market Power and Market Monitoring

285

participants and the market operator. Organizational factors therefore inﬂuence the effectiveness of the MMU and its ability to detect and prevent market abuses. They will also affect the effectiveness of enforcement activities following an investigation. These issues are examined next as the general ways in which markets monitors have been organized is described. There are more than 20 MMUs operating worldwide, most of which are listed in Table 7.2. A Market Monitors’ professional organization, the EISG, was formed to share market monitoring experiences and exchange information and ideas.37 An MMU may be established • • •

within an ISO: the MMU answers to the board of directors of the ISO; outside the ISO: the MMU in this case reports to a regulatory authority; or within a governmental agency that has full oversight and enforcement authority.

MMU established within the ISO: This model is followed by early restructured electricity markets in the United States,38 and more recently by the Southwest Power Pool (SPP). It was also initially adopted in the Canadian provinces of Ontario and Alberta, and in New Zealand.39 The US Federal Energy Regulatory Commission (FERC,) as part of its Standard Market Design in its Order 2000, required the creation of MMUs in each of the newly deregulated markets in the United States and emphasized the need for these units to be independent.40 All aspects of market monitoring are conducted by either the ISO or the Regional Transmission Organization (RTO) through these MMUs. The market monitor communicates with market participants, advises the board of directors, and reports to the board all suspicious activities and ﬁndings of market manipulation. The board may consist of members afﬁliated with market participants, or independent members, or a combination. The board may also have an independent market-monitoring advisor to assist its deliberations and decisions regarding issues identiﬁed by the MMU. In US markets under FERC jurisdiction, when the MMU identiﬁes a potential market manipulation or market power abuse, it is required to refer the case to FERC for further investigation. These MMUs have limited enforcement authority, but they can require that the market participant under investigation cease its potentially illegal or harmful activities. If the request is unsuccessful, FERC takes over the case as the agency that has full enforcement authority. One advantage of this model is that the MMU is more likely to operate with adequate resources and has immediate access to market operational data. The MMU staff can readily discuss market problems with ISO operators as they occur, which makes this the best model to ensure that the MMU has a good understanding of market operations and of the causes of market failures. The disadvantage of the model is that the MMU is less likely to be completely independent from the market participants or the ISO. Where the ISO’s

EBL

37 The EISG was created in 2000 and its members meet twice a year. The United Kingdom Ofﬁce of Gas and Electricity Markets’s (Ofgem’s) and the Scandinavian electricity market’s (NordPool’s) monitoring units are not member of the group at this time. A description of Ofgem can be found at: http://www.ofgem.gov.uk/ofgem/index.jsp. For a description of NordPool see: http://www.nordpool.com/ 38 These states are California, PJM, New York, and New England. 39 The Alberta and New Zealand markets have changed course and their MMUs are now outside the ISO and answer to government agencies or government appointed commissions. 40 See Federal Energy Regulatory Commission, Order 2000, (1999). Market monitoring was identiﬁed among eight core functions of newly established RTOs.

Competitive Electricity Markets

286

Table 7.2. Existing market monitoring units MMU within ISO, limited enforcement authority 1. Pennsylvania–New Jersey–Maryland (PJM)∗ 2. New York Independent System Operator (NYISO)† 3. Independent System Operator for New England (ISO–NE)‡ 4. California Independent System Operator (CAISO)§ with its Market Surveillance Committee (MSC)∗∗ 5. Southwest Power Pool (SPP)†† IMM outside of ISO 1. Midwest Independent System Operator (MISO)‡‡ 2. Market Surveillance Administrator, Alberta, in Canada§§ Government MMUs 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. ∗

Wholesale Market Oversight (WMO)∗∗∗ of the PUCT with its IMM located in ERCOT Australia Energy Regulator††† France Energy Regulation Commission‡‡‡ Japan Electric Power Exchange, Inc.§§§ Korea Electricity Commission∗∗∗∗ The Netherlands†††† The NordPool (Denmark, Finland, Norway, and Sweden)‡‡‡‡ Electricity Commission of New Zealand§§§§ Philippines Energy Regulatory Commission∗∗∗∗∗ Singapore Energy Market Company Pte. Ltd††††† United Kingdom Ofﬁce of Gas and Electricity Markets (Ofgem)‡‡‡‡‡ Independent Electricity System Operator (IESO)§§§§§ , Ontario, in Canada

EBL

Link to PJM Interconnection: http://www.pjm.com/about/pjm-links.html Link to NYISO: http://www.nyiso.com/public/index.jsp ‡ Link to ISO-NE: http://www.iso-ne.com/industry_links.html § Link to CAISO: http://www.caiso.com/surveillance/index.html ∗∗ Link to CAISO MSC: http://www.caiso.com/docs/2005/10/04/20051004104917266.html †† Link to SPP: http://www.spp.org/section.asp?pageID=17 ‡‡ Link to MISO: http://www.midwestiso.org/home §§ Link to Alberta Market Surveillance Administrator (MSA): www.albertamsa.ca ∗∗∗ Link to PUC WMO: http://www.puc.state.tx.us/wmo/about/aboutwmo.cfm ††† Link to Australia Energy Regulator: http://www.aer.gov.au/content/index.phtml/itemId/673322/fromItemId/651437 ‡‡‡ Link to France Energy Regulation Commission: http://www.cre.fr/marches/elec_gros.jsp §§§ Link to Japan Electric Power Exchange: http://www.jepx.org/ ∗∗∗∗ Link to South Korea Electricity Commission: http://english.mocie.go.kr/index.jsp and Power Exchange: http://210.223.85.230/english/ †††† Link to the Netherlands Ofﬁce of Energy Regulation: http://www.dte.nl/engels/about_dte/Links/Links.asp ‡‡‡‡ Link to Nord Pool: http://www.nordpool.com/ §§§§ Link to Electricity Commission of New Zealand: http://www.electricitycommission.govt.nz/rulesandregs/compliance ∗∗∗∗∗ Link to Philippines Energy Regulatory Commission: http://www.erc.gov.ph/ ††††† Link to Singapore Energy Market Company: http://www.emcsg.com/grids/emc_template_aq.asp?id=278&area=5¤tsection=27079 ‡‡‡‡‡ Link to United Kingdom Ofgem: http://www.ofgem.gov.uk/ofgem/index.jsp §§§§§ Link to Canada IESO: http://www.oeb.gov.on.ca/html/en/industryrelations/msp.htm †

Market Power and Market Monitoring

287

board of directors is a stakeholder board, the inﬂuence of the entities that are the subject of the MMU’s monitoring is more likely to be felt than where there is an independent board. MMU established outside the ISO: In the United States, this model has been adopted by the Midwest Independent System Operator (MISO). In MISO, all aspects of market monitoring are conducted by the MMU and are funded through the ISO budget subject to government approval. The MMU is selected in a competitive process conducted by the ISO and overseen by the local regulatory authority. While the MMU freely communicates with market participants, the ISO staff, and the members of the ISO’s board of directors, it does not answer to the board. All suspicious activities and ﬁndings of market manipulation are referred directly to the governmental authority for further investigation and enforcement. Compared to the previous model, the MMU similarly has limited enforcement authority but is independent from the ISO board. In Australia, the National Electricity Code Administration (NECA), which had market monitoring and enforcement authority until recently, was initially established by several states participating in the national electricity market as a completely separate entity from the ISO with its own board. In 2005, NECA was abolished and its market monitoring functions were taken over by a new government agency, the Australian Energy Regulator, which now places the Australian MMU in the next class. In Alberta, the Market Surveillance Administrator (MSA), which conducts market monitoring in that market, answered to the ISO board of directors until 2003. It became clear that there was a conﬂict of interest when the MSA had to investigate the behavior of the ISO. The MSA was then set up as an independent entity appointed by the government. At this time, the Alberta MSA is an independent entity with surveillance and investigative authority. The adjudicative and enforcement authority is placed with an independent hearing panel that is separate from the Energy and Utilities Board, the agency with all other energy-related regulatory responsibilities in the Province. In France, the Commission for the Regulation of Energy (CRE) has recently been charged with monitoring the wholesale electricity and gas markets by the French Parliament. When created in 2000, the CRE was given a structure that was intended to ensure complete independence from the market players as well as from the government. The budget of the CRE is a line item in the State’s budget and is approved by the Parliament. The agency has six government-appointed commissioners representing various sectors of the economy, an appointed president, and two appointed vice-presidents. To ensure their independence, CRE commissioners serve a 3- to 5-year term, cannot be dismissed, and are not eligible for a second term. The agency has full authority to collect and analyze market data and reports ﬁnding of potential violations to another independent government entity, the Competitive Council, vested with investigative and enforcement authority. The CRE is a fully independent entity that reports only to the French Parliament and the European Commission. Government MMUs: Market monitoring in this model is conducted from within a government agency with full investigation and enforcement authority. This appears to be a preferred model in many countries outside the United States, including New Zealand and Australia since 2005, and several European and Asian electricity markets. New Zealand started off with its MMU integrated within the ISO or market administrator as it is called in that country, but governmental authorities decided to switch the market surveillance team to a more independent setting. In 2003, the Electricity Commission was created to oversee the electricity market, and in 2004 the surveillance and compliance functions were entrusted to a Market Governance team that answers to the Commission. The six Commission members are appointed by the government.

EBL

288

Competitive Electricity Markets

A similar change occurred in Ontario where the market monitoring authority, the Market Surveillance Panel, was transferred in 2005 from being a panel of the ISO Board of Directors to become a panel of the provincial regulator, the Ontario Energy Board. This change was made to bolster the independence of the Panel; however, the importance of direct access to ISO data and operational expertise was recognized and the existing market oversight business unit within the ISO was contracted to continue to provide monitoring and analytic services to the Panel.41 ERCOT is in this class. The PUCT, which is the government agency responsible for market oversight and monitoring in Texas, has contracted with an entity to act as the IMM and conduct monitoring of the wholesale electricity market from ofﬁces located within the ISO facilities. The market monitoring functions in this case are funded by a fee charged to electricity customers and collected by the ISO, but the market monitor answers to the PUCT and refers all cases of possible market abuse and market rule violations to the PUCT for investigation and enforcement. This model has the potential to provide a high degree of independence from market participants and from the ISO. It offers the best opportunity for good coordination and a close working relationship between the monitoring and investigative arm and the enforcement arm of an agency with market monitoring responsibilities. However, in several markets where this general model is followed, large market players have launched lobbying efforts to inﬂuence legislators’ and regulatory bodies’ rulings, threatening the independence of investigations and enforcement decisions. Potential shortcomings of this model include the adequacy of resources and skills needed for effective market monitoring42 and reduced opportunity for interaction and coordination with the ISO operators. Thus, organizational factors can subject a market monitor to more or less pressure from the market actors whose behavior and operations it must scrutinize and report. Regardless of the type of organization selected, opportunities for a market participant to inﬂuence the outcome of an investigation exist, and the market monitor should be subject to strict ethics standards that address possible conﬂicts of interest. For example, an employee of the market monitor should not have a ﬁnancial interest in a power company that operates in the market it monitors, and should not accept employment with the company, or an afﬁliated company, or with the ISO, until a reasonable period of time after leaving employment with the MMU. Gifts or invitations should not be accepted. Market participant inﬂuence can also be exercised, sometimes in a subtle way, if there is unrestricted communication with the market monitor. MMU communication with market participants should be encouraged to the extent that the market monitor conducting an investigation needs access to market participant information and explanations of the facts. However, it should be structured to ensure no pressure is exercised, and no accidental disclosure of strategic information to the market participant occurs once the investigation becomes formal. Communication between the market monitor and the investigating and

EBL

41

This business unit, ring-fenced within the ISO, is also responsible for the enforcement of compliance with the market rules, a role unrelated to market power except for the specialized provisions related to the recovery of inappropriate congestion payments to market participants. 42 The experience in ERCOT and the Canadian Province of Alberta demonstrates that it is possible to effectively address this concern by establishing a funding mechanism, subject to government approval and control, in which an adequate portion of ISO-generated fees is used to fund market-monitoring activities.

Market Power and Market Monitoring

289

enforcement teams, on the other hand, should be encouraged and made as easy and unobstructed as possible. 7.3.4. MMU Objectiveness and Independence Not all the models described above have the necessary features to support the MMU’s goals of objectiveness and independence. It is clear that the market monitor must not be inﬂuenced by market participants, but it must also be independent from the market administrator or ISO. MMUs established within the ISO are more likely to experience interference from the organization they are supposed to monitor and evaluate if they have to report to the management of that organization, and if their budget is controlled by that organization. The Alberta MSA was reorganized and separated from the ISO when it found itself in the position of critically analyzing the behavior of the ISO. A more recent example of conﬂicting interests involving an MMU within an ISO was provided when the head of the PJM MMU reported to FERC that PJM management had interfered with his work and taken actions toward the MMU because of its dissatisfaction with the MMU’s analyses and conclusions.43 The Alberta solution was to set up the MMU as a stand-alone entity appointed by the government but not held to report to the ISO, the Utilities Board, a Board of Directors, or to the government, and with its own surveillance and investigative authority. In contrast, the PJM ISO has announced that it is considering disbanding its MMU and outsourcing the function, and it appears that the ISO would negotiate and administer the contract with the external MMU. This is not the case in Alberta where the MMU funding is also totally independent from the ISO, thereby leaving the MMU with no need to be wary of their treatment of the ISO in order to ensure a contract renewal. Entrusting the market monitoring function to an outside consulting ﬁrm raises some additional issues. Will the selected ﬁrm be inﬂuenced by its ﬁnancial relationship with energy companies that may operate or have parent companies that operate in the market being monitored? Will the staff assigned to market monitoring duties also work on unrelated projects for other companies, rather than devote its full attention to the task? Can a consultant working on a time-limited contract be as engaged in the work as a permanent staff member? How is continuity assured? Can an external ﬁrm working offsite acquire an intimate knowledge of the market’s operations necessary to understand the very subtle ways in which the market can be gamed? These issues, however, can be addressed to a large extent if a number of requirements are speciﬁed. For example, when the PUCT hired a consulting ﬁrm to perform the function of market monitor for ERCOT’s wholesale markets, the commission required that the director and its staff be dedicated to working uniquely as the ERCOT MMU, located onsite with their ofﬁces within the ISO, and reporting directly to the commission. The more fundamental issues that affect independence and objectivity are:

EBL

• • • 43

Who controls the budget of the MMU? Who has hiring and ﬁring authority over the director of the MMU? To whom is the MMU accountable?

See Electricity grid monitor claims interference. Wall Street Journal, 6 April (2007) More details about the charges are provided in PJM Market Monitor Joseph Bowring’s Prepared Statement presented at FERC’s technical conference on the Review of Market Monitoring Policies, Docket No. AD07-8-000, on April 5, 2007. Interested readers may follow the progress of this case at FERC.

290

Competitive Electricity Markets

The entity identiﬁed in response to these three questions must be as removed from the market participants and the market administrator the MMU evaluates as possible, and its decisions must not be inﬂuenced by them. The Alberta model of a stand-alone organization, which closely meets this description, seems to provide the best insurance for an objective and efﬁcient MMU that will effectively protect the competitiveness of the market. However, recent developments there indicate that it does not satisfy everyone. Financial stakes are high, and market participants that have been under investigation have been dissatisﬁed with a market monitor that they think has too much discretion and leaves them little opportunity for appealing enforcement decisions. As a result, a discussion is currently under way to bring more clarity to the standards of behavior the Alberta MSA uses as bases for its evaluations and compliance measures, as well as minimize the ability of the MSA to negotiate behavioral change. Such tradeoffs are not easy to make: overall these changes, if adopted, would reduce the level of discretion left to the MSA and may also reduce its effectiveness as a market monitor. 7.3.5. Evolution in market monitoring The role of MMUs has changed signiﬁcantly since wholesale power markets were ﬁrst open to competition.44 When MMUs ﬁrst took charge of their market monitoring responsibilities, there was no established way of conducting market monitoring, and the laws and rules of the market for the most part still had to be developed. More importantly, MMUs had no clear authority to investigate market manipulations, market power abuses, or other violations, and there was no case history. At that early stage, MMUs were often in a position to observe suspicious market activities but were not empowered to act on the undesirable behaviors they witnessed. MMUs have now entered a new phase. Rules and procedures are under development or have been developed with considerable input from stakeholders. The MMUs’ responsibilities are better deﬁned, and so are enforcement procedures. Most of the restructured electricity markets have put in place market participants’ behavior rules and have spelled out penalties for a range of prohibited activities.45 Over time, these rules have been reﬁned and expanded to reﬂect lessons learned. As a result, market participants have a better perception of the boundaries of acceptable behavior and they know what to expect if they step over those boundaries. MMUs, too, have a better sense of direction. The lines of responsibility between those who monitor and report market abuses, those who investigate, and those who enforce have been made clearer. The recent investigation of a large generation company in ERCOT and the market monitor’s ﬁnding that the company exercised its market power in violation of Texas laws and market rules are a good example of this evolution. The investigation took less than 3 months to ﬁnalize, and the PUCT staff was quick to issue a recommendation to levy a hefty ﬁne against the company based on the IMM’s investigation results. These recent developments provide a good example of the effectiveness that can be achieved when a market monitor is free of inﬂuence from market participants and works in full cooperation with enforcement authorities.

EBL

44

See Adib (2006). The Texas commission adopted an enforcement rule that spells out the duties and responsibilities of wholesale market participants and the procedures for an investigation. See PUCT Substantive Rule §25.503. 45

Market Power and Market Monitoring

291

Over time, electricity markets abroad have increasingly opted to establish their MMUs outside the market operator and in many instances within governmental agencies that have investigative and enforcement authority. Alberta in 2003, New Zealand in 2004, Ontario in 2005, and France in 2006 moved in that direction. In the United States, where the ﬁrst electricity markets to open to competition established their MMUs as part of the ISO, the MMU budget is still controlled by the ISO and MMU members are employees of the ISO. Even though the trend changed with MISO and ERCOT, where the separation between the ISO and the MMU became more prevalent, the ﬁrst US electricity markets to open to competition are showing increasing signs of conﬂictual relationships between MMUs and ISO management. Recent testimony presented at FERC by the head of the PJM MMU regarding the alleged PJM ISO interference with the work of the MMU shows that the issue of the market monitor’s independence has reached a critical point and public pressure is building up for FERC to address and correct such situations. However, it is desirable that any changes that bring about more market monitoring independence do so without sacriﬁcing access to crucial data sources. On the one hand, the monitoring function of the MMUs must remain within, or in close proximity to, the market operator, as the MMU’s work beneﬁts from close interactions with the market operator. On the other hand, the investigative, reporting, and enforcement functions of the MMUs must be insulated from possible interference not only from market participants, but also from pressures that may be exercised by the market operator or by a government. Establishing an MMU as a stand-alone organization is the organizational structure that is most likely to result in effective market monitoring. Placing an MMU within a government agency is also a positive development, but only to the extent that such agency is able to offer guaranties of independence in its investigations and enforcement decisions. Such organizational decisions are crucial in winning the public’s trust in electricity markets and achieving true competitiveness in those markets.

EBL

7.3.6. Examples of behavior leading to harmful market outcomes Market monitors are concerned with any action that causes a market price to be different than if market forces had prevailed. Some of the examples of market manipulation are the exercise of market power to affect prices, anti-competitive behavior affecting other market participants, strategic manipulation of schedules, and inadequate resource information that causes the ISO to make sub-optimal operational decisions in real-time. Further examples of harm to the market dealt with by MMUs across the world are brieﬂy presented in the Appendix 7A to this chapter. Market manipulation does not always involve an explicit rule violation. Operating guides and market rules cannot be exhaustively speciﬁc, and a market participant may be able to take advantage of ﬂaws in rules at the detriment of the market. Examples of gaming activities that do not necessarily violate market rules include scheduling resources in a way that increases the congestion payments received by an entity, and bidding strategies that pose no risk to the supplier but signiﬁcantly increase risk or costs for the rest of the market. In any case, the MMU must be on guard against any behavior that results in prices inconsistent with what would obtain under competitive conditions, regardless of whether the behavior was expressly prohibited or not.

292

Competitive Electricity Markets

7.4. Conclusions Market power is a fundamental problem that affects most electricity markets in transition from a regime of regulated monopolies to competition. Left unaddressed, restructuring will leave consumers at the mercy of unregulated monopolists. The problem is best addressed when legislators and regulators require structural changes and the large vertically integrated power companies inherited from the past divest their generation assets. If divestiture is not required, the market monitor could very well be the best bulwark against market power abuse and a pivotal defender of competition. If one or more generators participating in the newly deregulated market are so large that they can control prices in the absence of scarcity, regulators and market monitors will have the difﬁcult task of protecting electricity markets against market power abuses. To be effective, MMUs must preserve their independence. Organizational factors are important in determining whether market actors can exercise undue inﬂuence on the market monitor and the outcome of its investigations, and they can also help or hinder the market monitor’s access to the extensive and time-sensitive market information it needs to perform its duties. These factors have played out, lessons have been learned, and market monitoring has evolved over time. Around the world, MMUs are now better organized, and are receiving clearer directions. However, power companies have also stepped up their efforts to inﬂuence the political processes that affect market monitors. The burden is on legislative and regulatory authorities to ensure that market monitors are empowered to do their job effectively, that their independence is protected, and that the information they gather is utilized fairly but ﬁrmly to police electricity markets in transition to competition.

Acknowledgment

EBL

Parviz Adib and David Hurlbut were on the Wholesale Market Oversight staff of the Public Utility Commission of Texas from 2000 to 2007, and from 2002 to 2006, respectively. The opinions expressed in this chapter are those of the authors and do not represent the opinion of the Public Utility Commission of Texas. The authors especially would like to thank Danielle Jaussaud, Director of Market Analysis at the Public utility Commission of Texas for her invaluable feedback on various versions of this chapter. Our appreciation also goes to Dr Perry Sioshansi, President of Menlo Energy Economics, Keith Casey, Manager of Market Analysis & Mitigation, California Independent System Operator, Wayne Silk, Vice-President and COO, Alberta Market Surveillance Administrator, Harry Chandler, Director of Market Assessment and Compliance, Ontario Independent Electricity System Operator, and Nicolas Beaulaton, Aurelien Lecaille, and Marie Dufourg of the French Energy Regulation Commission for their excellent comments and suggestions. Finally, we would like to thanks Jess Totten, Director of Competition Division, Public Utility Commission of Texas, for reviewing an earlier version of this chapter.

7A. Appendix 7A.1. Some examples of harm to the market addressed by MMUS across the world Most of the forensic economics performed by market monitors falls into one of three categories: market power abuse, market manipulation, and market design inefﬁciencies. The ﬁrst involves suppliers that have market power, and their ability to control market outcomes. The second involves opportunistic behavior that does not necessarily require

Market Power and Market Monitoring

293

the ability to control prices. The third involves noncompliance with ISO rules in a way that is economically rational to the market players but poses reliability problems or leads to inefﬁcient market outcomes. While the following are real life examples, an attempt was made to avoid disclosing the names of markets or players involved in the behavior.46

7A.2. Market power abuse Examples of market power abuse fall under several categories: •

Price spikes. Every price spike – even those that occur when scarcity is evident – raises the specter of potential market power abuse. Market monitors are the “ﬁrst responders” called upon by regulators, market participants, and the public to investigate price excursions. Virtually every MMU has been called upon to investigate the price shocks resulting from a hurricane, an ice storm, a heat wave, or some other unusual event. • Economic withholding. A large supplier uses its market power to drive wholesale prices high. The supplier offers some energy at prices signiﬁcantly higher than the marginal cost of its units, and on many occasions these offers cause market prices to clear higher than they would have had the energy been offered closer to marginal cost. • Physical withholding. A supplier with most of its generating capacity located inside or near a load pocket decides to retire a number of older units. In one instance, the market monitor was called on to investigate whether the proposed plant retirements were economically justiﬁable, or whether they would increase the ability of the supplier to control prices in the load pocket. • Price mitigation. In some wholesale power markets, energy offers are mitigated when they fail the conduct and impact test. Market monitors may have to pay special attention to offers from suppliers that have market power due to the fact that their units must be committed to maintain reliability. In some markets, offers from such units are mitigated after the market monitor determines that there is economic withholding.

EBL

7A.3. Market manipulation Examples include: •

Local market power. A generation owner schedules its units day-ahead in such a way that it causes the market operator to forecast local congestion for the next operating day. To resolve the congestion, the ISO pays start-up costs for a unit owned by the same supplier and that is not scheduled to run. In this case, the supplier artiﬁcially creates congestion, then receives start-up payments for a unit that it would have run anyway. • Suspension of operations at power plants. Market protocols often require power plant owners to notify the market operator if it intends to cease operations at the plant. Failure to notify the ISO resulted in ﬁnancial penalties in at least one market. 46

For a description of speciﬁc gaming strategies and market manipulation in California and in Texas, see Market Oversight Division (2002a,b).

Competitive Electricity Markets

294 •

•

•

•

•

•

•

Failure to meet ancillary service obligations. One of the most common violations is the failure to meet ancillary service obligations. Generators receive payment to maintain spinning reserves, but then sell the associated energy when balancing energy prices rise. Such violations can cause reliability problems when the operator needs to deploy the reserves to maintain frequency. Insider trading. A market participant has information about the timing of a plant outage and uses that information to enter into ﬁnancial trades before news of the outage is released to the market. Price chasing. A generator puts un-scheduled energy into the grid when prices rise. The market monitor must determine whether the act is deliberate and has a ﬁnancial or reliability impact. Schedule manipulation. In one instance, several wind farm operators were receiving compensation for curtailing false generation schedules when the wind was not even blowing and there was no actual output to curtail. Failure to satisfy technical requirements. When a large generating company fails to meet the technical requirements for generator performance, it may result in a reliability event. In one market, two companies were ﬁned for such violations. Failure to justify unit inﬂexibility. A generating company declares that its resources’ higher operating limits are equal to its lower operating limits without a proper justiﬁcation, making the unit unavailable to be dispatched by the market operator for reliability purposes. Invalid gate closures. A generator changes its offer after the market closes without proper justiﬁcation. In one market, the market monitor found that a number of breaches of that nature were committed within a short period of time, and that procedures that could have prevented the breaches from occurring were not in place.

EBL

7A.4. Market design deﬁciencies •

Reference price creep. In markets that use price or offer mitigation tools based on reference prices such as the Automatic Mitigation Procedure (AMD) of the New York and New England markets, the methodology used for calculating reference prices can over time yield results that diverge signiﬁcantly from marginal costs. In one case, the market monitor found that resources that do not face competition, such as units located in load pockets, resources needed to solve transmission congestion, or reliability-must-run resources skewed the reference prices, requiring a change in the reference price calculation methodology. • Morning and evening ramps. At the beginning and at the end of the standard 16-hour peak period each day, a large number of generators simultaneously ramping up in the morning and ramping down in the evening can result in frequency deviations requiring large deployments of regulation reserves to maintain power balance. Market monitors’ recommendations to smooth out these ramping problems have not been easily accepted by the market. • Imputed value for out-of-merit actions. When a market operator relies on out-of-market actions to create additional operating reserves for reliability purposes, these actions have a market impact that is not recognized. Market monitors have suggested that the market operator impute a value for these actions that would be reﬂected in the energy price.

Market Power and Market Monitoring

295

7A.5. Improvements in system operators’ performance •

Improvement in ISO load forecast. An evaluation of the ISO’s load forecast methodology by the market monitor may be useful in improving load forecast accuracy and reducing the balancing energy and regulation deployments caused by ﬂawed load forecasting methodologies. • Improvement in ISO procurement practices. One market monitor conducted an investigation into the practices of the market operator in the procurement of reliabilitymust-run services and concluded that the market operator had abused its position as monopoly buyer in its negotiations with transmission service providers. The market monitor’s recommendations resulted in improved procurement procedures and a clariﬁcation of the ISO’s authority relative to its procurement requirements. No ﬁnancial penalties were applied but the ISO’s procurement practices were improved going forward.

References Adib, P. (2006). Discussion by Parviz Adib Restructuring Today, Market Monitoring: Is the Focus changing? Audio Conference, Washington, D.C., 28 July. Armstrong, M. and Sappington, D. (2006). Regulation, competition, and liberalization. J. of Econ. Lit., 44(2), 325–66. Baumol, W.J. (1982). Contestable markets: An uprising in the theory of industry structure. Am. Econ. Rev., 72, 1–15. Borenstein, B. (2002). The trouble with electricity markets: Understanding California’s restructuring disaster. J. of Econ. Pers., 16, 191–212. Borenstein, B., Bushnell, J., and Wolak, F. (2002). Measuring market inefﬁciencies in California’s restructured wholesale electricity market. Am. Econ. Rev., 92, 1376–1405. Falk, J. (2004). The social beneﬁt of the limited exercise of local market power. Elec. J., 12–23. Federal Energy Regulatory Commission (1999). Commission acts to assure independence of PJM market monitoring. New release. 20 September Kahn, A., Cramton, P.C., Porter, R.H., and Tabors, R.D. (2001). Uniform pricing or pay-as-bid pricing: A dilemma for California and beyond. The Elec. J., 14, 70–9. Leveque, F. (2006). Antitrust enforcement in the electricity and gas industries: Problems and solutions for the EU. The Elec. J., 19, 27–34. Market Oversight Division (2002a). Enron’s wholesale power trading strategies in California: How did they work? Can they happen in the Electric Reliability Council of Texas? Prepared for the Texas Legislative Oversight Committee, Public Utility Commission of Texas, 18 June, Austin, Texas. Market Oversight Division (2002b). Mitigation measures for gaming opportunities in ERCOT wholesale electricity market. A Presentation to the Texas Legislative Oversight Committee, Public Utility Commission of Texas, 18 June, Austin, Texas. Moss, D. (2005). Electricity and market power: Current issues for restructuring markets (a survey). American Antitrust Institute working paper No. 05-01, February. Potomac Economics (2005). Investigation into the causes for the shortages of energy in the ERCOT balancing energy market and into the wholesale market activities of TXU from October 27 to December 8, 2004, April, Austin, Texas. Potomac Economics (2007). Investigation of the Wholesale Market Activities of TXU from June 1 to September 30, 2005. Independent Market Monitor, March. The report was futher revised in September 2007. Restructuring Today (2006). Market monitoring: Is the focus changing? Audio Conference, Washington, D.C., 28 July. Shepherd, W.G. (1984). Contestability vs. competition. Am. Econ. Rev., 74, 572–87. Stoft, S. (2002). Power System Economics. New York: John Wiley & Sons.

EBL

296

Competitive Electricity Markets

Sweeney, J.L. (2006). California electricity restructuring, the crisis, and its aftermath. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Elsevier. Texas Public Utility Regulatory Act (2005). Texas Public Utility Regulatory Act, as amended in 2005, Title II, Texas Utilities Code. US Department of Justice and US Federal Trade Commission (1997). Horizontal merger guidelines, revised in 1997. US Department of Justice (2006). Commentary on the horizontal merger guidelines, March. Wolak, F. (2004). Lessons from international experience with electricity market monitoring. Center for the Study of Energy Markets, University of California Energy Institute, CSEM WP 134, June. Available at: http://www.ucei.berkeley.edu/PDF/csemwp134.pdf.

EBL

Chapter 8 Demand Participation in Restructured Markets JAY ZARNIKAU Frontier Associates LLC and LBJ School of Public Affairs, The University of Texas at Austin, Texas, USA

Summary1 Demand response is of particular importance in restructured markets that face some unique challenges in maintaining resource adequacy and preventing market power abuse. By responding to prices, by providing an ancillary service, or by offering to curtail usage in response to an interruption request, the participation of consumers or loads can contribute to the efﬁcient operation of an electricity market. Establishing the policies and market structure that will enable demand participation has proven difﬁcult to date. Nonetheless, advances in technology and reductions in infrastructure costs hold the potential to raise demand participation to new heights.

EBL

8.1. Why Demand-Side Participation is Important in Restructured Markets In most markets, there is little need to make demand more elastic. In competitive markets for most goods and services, retail prices adjust to reﬂect the relative abundance or scarcity of the good or service and changes in the cost of providing it to the market, and consumers adjust their purchasing behavior accordingly. But inelastic demand has been and remains a feature of electricity markets. In the 1980s and 1990s, the establishment of real-time pricing programs sought to address this problem by providing consumers with prices designed to mimic competitive prices that would reﬂect short-run marginal costs (Barbose et al., 2004). More recently, critical peak pricing programs became a popular means of eliciting a response from consumers by providing strong price signals in a limited number of high-cost or high-demand periods. Yet, aside from these exceptions, regulated retail prices seldom reﬂected the frequent changes in the value of electricity generated, transmitted, and distributed by the utility 1 Fereidoon P. Sioshansi, Parviz Adib, Alison Silverstein, Charles Goldman, and Nat Treadway provided valuable comments on earlier drafts of this chapter.

297

298

Competitive Electricity Markets

system and sold to an ultimate retail customer. Tariffs reﬂected the long-term average cost of providing electricity to a class of consumers. For smaller electricity consumers, the costs of metering energy usage at less than monthly intervals thwarted pricing schemes to reﬂect changes in the value of electricity in retail prices in hourly or 15 minute intervals on a wide-scale basis (see, e.g., Zarnikau et al., 1990; O’Sheasy, 2003). Load management programs and other types of “peak clipping” or “load shifting” demand-side management sought to elicit responses from consumers in situations where pricing programs might not prove effective. This includes situations where the costs of metering and settling smaller loads based on real-time prices might outweigh the beneﬁts. While demand-side participation is important in regulated markets, the response of consumers or “loads” to price signals is even more crucial to the efﬁcient operation of restructured power markets. As noted by US Federal Energy Regulatory Commission (FERC): “Demand response is essential in competitive markets, to assure the efﬁcient interaction of supply and demand, as a check on supplier and locational market power, and as an opportunity for choice by wholesale and end-use customers” (Federal Energy Regulatory Commission, 2002). The National Association of Regulatory Utility Commissioners (NARUC) has called for regulatory commissions to accommodate demand-side resources and “remove any unnecessary barriers to customer responses to such wholesale market price signals” (NARUC, 2000). Through the Energy Policy Act of 2005, the US Congress afﬁrmed the importance of expanding demand response opportunities as a matter of national policy, stating: It is the policy of the United States that time-based pricing and other forms of demand response, whereby electricity customers are provided with electricity price signals and the ability to beneﬁt by responding to them, shall be encouraged, the deployment of such technology and devices that enable electricity customers to participate in such pricing and demand response systems shall be facilitated, and unnecessary barriers to demand response participation in energy, capacity and ancillary service markets shall be eliminated (Energy Policy Act of 2005, Section 1252(f); see also US Department of Energy, 2006).

EBL

One factor cited as responsible for the collapse of California’s competitive electricity market in 2000–01 was an absence or inadequacy of demand response (Faruqui et al., 2001; see also Sweeney, 2006). Demand response encompasses changes in consumer electricity consumption decisions in response to changes in the price of electricity as well as programs that require consumers to curtail their usage at the request of an independent system operator (ISO), utility, or other authority in return for a price discount or payment. The US Demand Response Coordinating Council offers a formal deﬁnition: Providing electricity customers in both retail and wholesale markets with a choice whereby they can respond to dynamic or time-based prices or other types of incentives by reducing and/or shifting usage, particularly during peak periods, such that these demand modiﬁcations can address issues such as pricing, reliability, emergency response, and infrastructure planning, operation, and deferral.2

2

http://www.demandresponseinfo.org/id46.htm

Demand Participation in Restructured Markets

299

The US Department of Energy (DOE) similarly deﬁnes demand response as: Changes in electric usage by end-use customers from their normal consumption patterns in response to changes in the price of electricity over time, or to incentive payments designed to induce lower electricity use at time of high wholesale market prices or when system reliability is jeopardized (US DOE, 2006). Many of these deﬁnitions of demand response encompass the load management and innovative pricing concepts that attained popularity under regulation. In markets with relaxed regulatory oversight, demand response can restrain prices to economically efﬁcient levels (Rosenzweig et al., 2003). A small amount of demand response can yield signiﬁcant reductions in short-term wholesale electricity prices (Caves et al., 2000; Faruqui and George, 2002). For example, recent analyses by the staff of the New England ISO suggests that a 500 MW increase in demand response participation would reduce wholesale costs by $32 million annually (ISO-New England, 2006). A 3% load reduction in the top 100 hours in ﬁve mid-Atlantic zones could yield total annual beneﬁts of from $138 to $281 million (Brattle Group, 2007). Demand response becomes all the more important when it is recognized that most restructured electricity markets are far from “perfectly competitive.” Thus a reduction in demand in response to a price spike is crucial to constrain the ability of suppliers to raise prices to inefﬁcient levels (i.e., price levels inconsistent with consumer willingness or ability to pay), due to the exercise of market power or other anticompetitive behaviors. Insufﬁcient demand response is sometimes used as a justiﬁcation for wholesale price caps (which in turn may dampen or jeopardize price response. In markets where an “energy-only” approach is adopted to maintain resource adequacy, demand response may play an important role in maintaining a balance between supply and demand. This is particularly important in light of the cyclical nature of power plant construction activity. During the periods when a market is left with inadequate reserve margins, demand response can provide an important backstop (Parviz et al., this volume-b). It is sometimes argued that demand-side resources can be used to defer or displace transmission investments in either a regulated or a competitive market. There have been few situations in which this potential beneﬁt has been successfully exploited (FERC, 2006). However, some encouraging programs have been launched in California, New York, and the Paciﬁc Northwest region of the United States.3 Carefully crafted demand response programs can be used to foster reliability in realtime system operations. High wholesale prices or participation in programs through which loads curtail in response to instructions from a system operator in return for some ﬁnancial compensation can assist the system operator in balancing supply and demand in real-time and in managing reliability during emergency conditions. Finally, when viewed as a call option, demand response may provide a variety of risk management beneﬁts to an ISO or load-serving entity in a competitive market.4 As electricity markets are redesigned to facilitate wholesale and/or retail competition, stakeholders and policymakers face the challenge of ensuring that consumers are presented with accurate price or curtailment signals and the appropriate incentives to react to those signals. Yet, responses by retail energy consumers to price changes and the introduction of demand-side resources (e.g., interruptible or curtailable loads) into competitive markets

EBL

3 4

See FERC (2006), pp. 115–7. See FERC (2006), pp. 11–2.

300

Competitive Electricity Markets

primarily designed for power plants pose certain technical and economic challenges. Key policy questions include the following: •

•

• •

• • • • •

Must demand-side participation be fostered with special programs, or can a means of promoting demand-side participation through a competitive market structure be found? To what extent can demand response be relied upon to match supply and demand, especially in markets that plan to rely upon “energy-only” resource adequacy mechanisms? How will changes in consumption in reaction to wholesale prices affect the need for generation and ancillary services? How should a consumer’s need for sufﬁcient time to respond to prices be balanced against an ISO’s desire to reduce forecast errors (to increase the accuracy of prices) in near-real-time operations? Can demand-side resources provide the same types of ancillary services as supplyside resources (generators)? What metering policies are needed in order to facilitate beneﬁcial demand response? Will new technologies come to the rescue? Can demand-side resources play a role in markets for transmission rights? Is a “quasi-LMP” market structure, in which power plants face nodal prices and loads face a different set of zonal prices, sustainable?

While these policy questions are difﬁcult, the stakes are high, and the opportunities to improve electricity markets through effective demand response efforts are enormous. Indeed, efﬁcient demand response can contribute to resolving many of the problems associated with market power, resource adequacy, and reliability discussed in other chapters of this volume. This chapter reviews the challenges associated with facilitating demand-side participation in competitive markets for power, provides a summary of various demand-side initiatives in restructured markets in North America, and contributes a case study describing some of the problems faced in promoting demand-side participation in the Electric Reliability Council of Texas (ERCOT) market, often cited as the most successful of North America’s restructured retail markets.

EBL

8.2. Barriers and Opportunities In any electrical network – regulated or restructured – the costs associated with energy generation, transmission, distribution, and meeting various operational constraints change continuously over time. The costs associated with serving customers in different areas of a network may vary greatly. Yet, electricity consumers have traditionally faced ﬂat systemwide rates, which seldom varied over time or space within a utility service area. Economic efﬁciency requires that prices bear some relationship to marginal costs, a condition that was seldom satisﬁed under traditional ratemaking. Successful demand response requires a correct combination of customer characteristics, economic incentives, metering and communications technology, market design, and policy. Fundamentally, consumers need the motivation and means to respond to price signals or indicators of reliability problems (Hirst, 2002, p. 2). Certain loads (e.g., municipal wastewater pumping, pipelines, and steel mills) tend to possess considerable ﬂexibility in their operations and are natural candidates for nearly any demand response initiative. Production at such facilities may be delayed during a high-price period at relatively little

Demand Participation in Restructured Markets

301

economic loss to the facility. Similarly, water heating can be deferred for a few hours with no noticeable loss of comfort to a residential consumer. Technologies capable of enabling demand response may include backup generation, control systems, load monitoring equipment, and energy storage devices. High penalties for failure to comply with curtailment requests and uncertain payments tend to discourage participation in voluntary curtailment programs (Neenan et al., 2002). Response may be affected by incentive levels, notice periods, the importance of energy costs to the consumer, the communications infrastructure, the customer’s sophistication in energy consumption decisions, and a multitude of other customer-speciﬁc factors. In a regulated market, a vertically integrated utility may realize a variety of beneﬁts from the ability to control or reduce loads. The cost of operating expensive peaking capacity may be reduced. Generating capacity additions may be deferred. In some rare situations, transmission or distribution capacity additions could be deferred. Overall system reliability may beneﬁt. In regulated markets, demand-side participation can be fostered by utilities and regulatory authorities through the establishment of interruptible tariffs and demandside management (e.g., load management) programs.5 However, in an unbundled market, there are some unique economic challenges associated with demand-side participation. Different entities may receive value from different segmented beneﬁts. For example, a retailer could beneﬁt from lower purchased power costs or the ability to better manage power procurement risks, a transmission and distribution services provider might be the beneﬁciary of deferred transmission or distribution improvements, and the ISO might beneﬁt from an additional tool to maintain reliability. But, no single entity (i.e., the retailer, wires provider, or ISO) is likely to derive sufﬁcient beneﬁts within its own sphere of operations to justify the establishment of a program. The “value chain” has been severed and economies of scale and scope may be lost.6 In competitive markets, the design of the market rules must be crafted to accommodate and facilitate demand-side participation. Alternatively, special demand response programs may be grafted onto the competitive market. For a retailer, the possibility of stranded costs, which may result if the customer later switches suppliers, may discourage investments in metering, control technologies, and technical assistance necessary to implement a successful program. This is particularly true for residential direct load control, where program equipment infrastructure costs can be quite signiﬁcant relative to the economic value of the demand reduction that can be realized. While third-party demand reduction aggregators can play a valuable role in facilitating demand response in competitive markets, they face a number of barriers. Access to the meter may be an obstacle. Who owns the meter? Is it just a “cash register” for the retailer, or can it also be used by a third party who is developing a demand response program? In some markets, retailers are concerned about energy services companies (ESCOs) and demand aggregators “interfering” with their relationship with their customer, particularly if demand response may have an impact on the retailer’s procurement of generation. Settlement procedures may be complicated when one entity is selling generation to a retail consumer while another entity curtails or shifts the temporal pattern of the electricity that is consumed. Reliable real-time communications between system operators and retail customers pose another obstacle. It has been noted that currently “the protocols for sending the signal that

EBL

5 It is sometimes argued that load management programs are not true demand response if their deployment is not sufﬁciently driven by prices or market conditions. 6 An excellent discussion of this is provided in Spine (2002).

302

Competitive Electricity Markets

capacity is tight and voluntary load shedding is needed is time-consuming, error-prone, and mostly manual” (Neumann et al., 2006). Further, the feedback from the customer back to the system operator may be insufﬁcient to assure the operator that adequate demand response has been achieved in real-time. Customer education may also be needed. Some energy consumers may lack the knowledge and sophistication necessary to take advantage of demand response opportunities. They may not understand the load associated with various equipment, the ﬂexibility that they may have with respect to equipment operation, and electricity market economics and opportunities. Another challenge is political. Generation owners recognize the ability of demand response to reduce the value of generation assets, given the potential of demand response to constrain prices and provide a substitute for certain ancillary services and planning reserves. Generators often wield considerable power within the stakeholder processes used to design and operate markets. Consumers who could provide demand-side resources have primary interests that lie outside of power markets. Consequently, demand response initiatives may be difﬁcult to establish when a stakeholder process is used to design a competitive market. “Consumer segments” may be less united and powerful than other market participant interests, and have relatively less understanding of the energy business than power producers. Demand response poses some interesting challenges for market monitors. In some situations, it may be difﬁcult to distinguish “gaming” from demand response. For example, if a scheduling entity or market participant informs the ISO that its supply will exactly equal demand, but, after seeing a high price, curtails load under its control in order to create a net injection into the market and earn a proﬁt from the unscheduled net sale to the market, is the market participant guilty of any improper behavior? The answer should be no. Yet there may be some suspicion that the (net) supply available from the entity was not as limited as the market was led to believe. It is important that market monitors have the information and tools necessary to distinguish between these situations. For ISOs, demand response can introduce some complications into the operation of power systems. In the absence of a priority pricing scheme (Oren et al., 1986), where the prices at which load will curtail are announced to the market, the ISO may have difﬁculty understanding the slope or elasticity of the market’s aggregate demand curve in a realtime energy market. ISOs sometimes complain that demand response affects the accuracy of their near-term forecasts of demand, thus complicating the task of balancing demand and supply in real-time. If more generation is scheduled than needed, additional costs may be imposed upon the market. The usual problems associated with setting customer baselines and estimating the amount of demand reduction achieved through a demand response action can become more contentious in a competitive market setting, where multiple market participants (rather than a single vertically integrated utility) have a stake in the method of calculating demand response and in designing the formulas used for compensation. In nodal or locational marginal cost (LMP) markets7 further difﬁculties may arise when the power consumed by an energy consumer is settled on a zonal basis while generation resources are settled on a nodal basis. Customers served by the load-serving entity will face average or muted zonal prices that will fail to reﬂect the true value of electricity consumed at a node or the value of a reduction in demand at a node. While the theory behind

EBL

7

See Chapter 4 by Singh (this volume).

Demand Participation in Restructured Markets

303

LMP markets clearly requires loads to be settled at nodal prices (Baughman et al., 1997; Schweppe et al., 1988), equity and technical concerns have resulted in compromises. It is difﬁcult to charge some consumers more than other consumers for transmission congestion charges when they historically paid the same rates toward transmission investments to the same vertically integrated utility prior to restructuring. Externalities may be a problem – the actions of your neighbor on the transmission network could affect your transmission costs. There may also be practical difﬁculties in calculating and communicating nodespeciﬁc prices to consumers and in the settlement process. Finally, it should be noted that wholesale or retail rate caps may dampen price signals and response. Ironically, the concern over insufﬁcient price elasticity that prompts policymakers to impose price caps reduces the incentives and interest by consumers in responding to prices. In addition, there are many regulatory barriers to be overcome by demand response programs. In a recent report, FERC (2006) identiﬁed nine “regulatory barriers” to demand response, as presented in Table 8.1. Many of these appear to be applicable to demand response initiatives outside of the United States as well.

Table 8.1. Regulatory barriers for demand response identiﬁed by FERC (2006) 1. Disconnect between retail pricing and wholesale markets. Retail rates for most customers are ﬁxed, while wholesale prices ﬂuctuate. 2. Utility disincentives associated with offering demand response. Reductions in customer demand reduce utility revenue. Without regulatory incentives such as rate decoupling or similar incentives, electric utilities lack an incentive to use or support demand response.

EBL

3. Cost recovery and incentives for enabling technologies. Utilities are reluctant to undertake investments in enabling technologies such as advanced metering unless the business case and regulatory support for deployment is sufﬁciently positive to justify the outlay. These investments may require an increase in rates. It is uncertain whether and how would regulators allow these costs to be recovered. 4. The need for additional research on cost-effectiveness and measurement of reductions. There are deﬁciencies in the measurement of demand response and assessment of its cost-effectiveness. Cost-effectiveness tests that have been used by regulators must be improved to reﬂect changes in the industry, especially in organized markets. 5. The existence of speciﬁc state-level barriers to greater demand response. Policies of retail rate regulators and state statutes in several states have created barriers to implementing greater levels of demand response, especially by exposing customers to time-based rates. 6. Speciﬁc retail and wholesale rules that limit demand response. Certain wholesale and retail market designs have rules and procedures that are not conducive to demand participation. 7. Barriers to providing demand response services by third parties. Shifting regulatory rules that allow third parties to provide demand response and potential sunset of various demand response programs are a disincentive to demand response providers. Because third parties often bear the risks of programs dependent on enabling technologies, they need long-term regulatory assurance or longterm contracts to raise the capital needed to invest in enabling technologies. 8. Insufﬁcient market transparency and access to data. Lack of third-party access to data has been identiﬁed as a barrier to demand response. A related but more fundamental barrier related to data is timely access to meter data. 9. Better coordination of federal-state jurisdiction affecting demand response. While states have primary jurisdiction over retail demand response, demand response plays a role in wholesale markets under Commission (FERC) jurisdiction. Greater clarity and coordination between wholesale and state programs is needed.

Competitive Electricity Markets

304

While the challenges associated with promoting demand-side participation in a competitive market may be formidable, advances in technology are making some aspects of it more practical. Metering technology has advanced and the cost of metering loads in nearreal-time has declined. An increasing number of utilities are now adopting or considering advanced system-wide smart meter technologies. With an advanced metering infrastructure, consumption of smaller customers can be more easily matched with prices, and the amount of demand reduction obtained from small consumers can be measured directly from interval metering data rather than through statistical sampling techniques. The Internet, under-utilized paging networks, and other communications media also open up new avenues for communicating electricity prices to consumers. Finally, systems designed to control or optimize the operation of equipment at industrial facilities, commercial buildings, and homes, and other enabling technologies have improved, expanding the horizon for more effective demand responses in the near future.

8.3. Demand-Side Participation: From Traditional Utility Efforts to Restructured Markets In a traditional regulated utility setting, a very wide variety of programs and tariffs were established to improve customer exposure to price signals and provide demand reduction at the request of the utility. Such efforts include8 : •

•

•

•

•

•

Direct load control demand-side management programs, involving the control of customer appliances or equipment (e.g., air-conditioners, water heaters, or pool pumps) from a central location. The program participant receives an incentive payment or bill discount. Curtailment programs, providing a ﬁnancial incentive to participants who agree to control their electricity-intensive equipment so as to reduce their electrical usage at the request of the utility. Interruptible tariffs, providing electricity at a discounted price to large industrial or commercial consumers who agree to interrupt their purchases at the request of the utility during a supply shortage or instantaneously in response to a system emergency. Time-of-use pricing, charging different retail prices for electricity purchased during different blocks of time. Typically, different periods within a day receive different prices. The pricing periods and price differentials are predetermined to correspond with average historical price patterns. Critical peak pricing, permitting retail prices to reﬂect their true market values or provide a strong price signal a few times a year. Otherwise, prices will often follow a time-of-use pattern. Real-time pricing, changing prices on an hourly basis to reﬂect the cost of providing electricity to the consumer each hour. Customer baselines and other features are often applied to ensure the utility’s revenue neutrality with respect to the consumer’s historical usage pattern (E Source, 1995).

EBL

These options are summarized in Table 8.2. 8

See also US DOE (2006).

Demand Participation in Restructured Markets

305

These various products and services may be differentiated by: •

Whether the customer is exposed to a price signal (economic or market-based demand response) or curtailment requests by a utility (reliability- or contingencybased programs). • Advance notice periods, which may range from instantaneous response (interruptible service involving under-frequency relays) to day-ahead response (e.g., participation in a day-ahead energy market). • Type of incentive provided to the participant, which may be a direct payment, a discounted electricity price, or time-based rates. DOE (2006) and FERC (2006) draw a distinction between “incentive-based demand response” and “time-based rates.” • The degree of commitment or cessation of control provided by the participant, which may range from third-party control over the operation of customer-owned equipment (e.g., direct load control) to voluntary actions (in the case of most price response programs). There are many similarities among the challenges and opportunities associated with promoting demand-side participation in markets with different structures. Yet, there are some differences as well. Regulated tariffs are available in regulated markets. Regulators often deviate from market economics to design pricing structures that meet long-term societal goals. There tends to be greater certainty associated with the costs and beneﬁts of participating in regulated programs or accepting service under regulated tariffs. In restructured markets, the forces of supply and demand, retailer pricing strategies, and stakeholder processes tend to shape opportunities for demand-side participation. In a restructured competitive market, a load-serving entity might offer the same types of programs and pricing strategies as might be offered in a traditional regulated setting in order to shape its generation requirements. However, a variety of additional services might be offered, such as:

EBL

•

Direct access to a competitive wholesale market, providing the consumer with direct exposure to real-time or day-ahead market prices. • Demand bidding, allowing the consumer to submit a formal offer to curtail its electricity usage to a day-ahead or real-time market and receive a market price for the demand reduction. • Direct participation in ancillary services markets, providing an interruptible energy consumer with an opportunity to provide an operating reserve. • Participation in Installed Capacity or ICAP markets, paying an interruptible load an incentive in return for providing a call option that an ISO can exercise in order to curtail the load in the event of a system emergency. Whether these services or opportunities are offered by a traditional vertically integrated utility, a retailer, or the ISO or regional transmission organization (RTO) responsible for ensuring reliability may depend upon the market structure. Of particular note are the reliability-focused programs operated by ISOs in the northeastern United States and in California.9 The New England ISO (ISO-NE), the New York ISO (NYISO), and the PJM ISO (serving a large region including Pennsylvania, New Jersey, Maryland, and the upper Midwest) have all established successful programs focused on reliability. Reliabilityfocused programs tend to be reﬂected in reserve margin calculations and system planning 9

For a survey of programs operated by utilities, see FERC (2006).

306

Competitive Electricity Markets

Table 8.2. Demand response options Price-based options

Incentive-based programs

Time-of-use (TOU): a rate with different unit prices for usage during different blocks of time, usually deﬁned for a 24-hour day. TOU rates reﬂect the average cost of generating and delivering power during those time periods.

Direct load control: a program by which the program operator remotely shuts down or cycles a customer’s electrical equipment (e.g., air-conditioner, water heater) on short notice. Direct load control programs are primarily offered to residential or small commercial customers. Interruptible/curtailable (I/C) service: curtailment options integrated into retail tariffs that provide a rate discount or bill credit for agreeing to reduce load during system contingencies. Penalties may be assessed for failure to curtail. Interruptible programs have traditionally been offered only to the largest industrial customers.

Real-time pricing (RTP): a rate in which the price for electricity typically ﬂuctuates hourly, reﬂecting changes in the wholesale price of electricity. Customers are typically notiﬁed of RTP prices on a day-ahead or hour-ahead basis.

Demand bidding/buyback programs: customers offer bids to curtail based on wholesale electricity market prices or an equivalent. Mainly offered to large customers [e.g., one megawatt (MW) and over].

Critical peak pricing (CPP): CPP rates are a hybrid of the TOU and RTP design. The basic rate structure is TOU. However, provision is made for replacing the normal peak price with a much higher CPP event price under speciﬁed trigger conditions (e.g., when system reliability is compromised or supply prices are very high).

Emergency demand response programs: programs that provide incentive payments to customers for load reductions during periods when reserve shortfalls arise.

Capacity market programs: customers offer load curtailments as system capacity to replace conventional generation or delivery resources. Customers typically receive dayof notice of events. Incentives usually consist of upfront reservation payments, and face penalties for failure to curtail when called upon to do so.

EBL

Ancillary services market programs: customers bid load curtailments in ISO/RTO markets as operating reserves. If their bids are accepted, they are paid the market price for committing to be on standby. If their load curtailments are needed, they are called by the ISO/RTO, and may be paid the spot market energy price.

Source: US Department of Energy (2006). Beneﬁts of demand response in electricity markets and recommendations for achieving them: A report to the U.S. Congress pursuant to Section 1252 of the Energy Policy Act of 2005. February.

studies, while economic programs tend to be viewed as less reliable as a source of planning resources. Some prominent reliability-focused ISO-level demand response opportunities in US markets are summarized in Table 8.3. ISO-NE offers reliability programs with notice periods of 30 minutes or 2 hours, in addition to a real-time program. In 2006, 641 MW of load was available for interruption

Demand Participation in Restructured Markets

307

Table 8.3. Reliability-focused programs offered by selected ISOs in the United States ISO

NY ISO

Program

Number of participants

Potential demand Reduction (MW)

Emergency demand response program ICAP/special case resources

810

369

1460

861

ISO-NE

Emergency load response

1490

801

PJM

Emergency load response program

4427

1081

Cal-ISO

Various programs operated by utilities

NA

1800

Sources: Demand Response Department of ISO New England (2006).ISO New England/NEPOOL demand response working group meeting. Presentation, 6 December; NYISO (2006). November 2006 demand response registration. Presentation by Price Responsive Load Working Group, 8 December; PJM (2006). Load response activity report, January through September 2006. 30 September; and e-mail correspondence from Greg Fishman of the California ISO to Rebecca Farrell, 1 December 2006.

through the 30 minute program (in an ISO with a peak demand of 28 127 MW).10 An additional 117 MW is enrolled in the real-time program and 26 MW was available through the 2 hour program. Participants in the 30 minute program receive the higher of the realtime zonal price or $500/MWh for a minimum of 2 hours, while participants in the 2 hour notice program earn the higher of the real-time zonal price or $350/MWh for a minimum of 2 hours. Real-time resources receive the higher of the real-time zonal price or $100 per MWh. The majority of the 30 minute notice demand response is in Connecticut, where a supplemental program provides up to an additional $150/kW-year in this transmissionconstrained area. Participants in these ISO-NE programs qualify to participate in this ISO’s market for installed capacity (ICAP) planning reserves. The ICAP market is an important revenue stream for program participants. The value of ICAPs in the New England market was $3.05/kW-month in December 2006.11 In the NYISO, two major reliability-focused programs are available: the Emergency Demand Response Program (EDRP) and the ICAP/Special Case Resources (SCR) program. In the EDRP program, 161 MW was available in New York City and Long Island in 2006, while 208 MW was available in the rest of the state.12 Most of the EDRP resource is

EBL

10

Demand Response Department of ISO New England (2006). ISO New England/NEPOOL demand response working group meeting.Presentation, 6 December. 11 Figures in this paragraph are from: ISO-NE, FERC Electric Tariff No. 3, Section III, Market Rule 1 – Standard Market Design, p. 263, at: http://www.iso-ne.org/regulatory/tariff/sect_3/ section_iii_mr1_standard_mkt_design.pdf. 12 Figures in this paragraph are from: NYISO (2006).November 2006 demand response registration. Presentation by Price Responsive Load Working Group, 8 December, available at: http:// www.nyiso.com/public/webdocs/committees/bic_prlwg/meeting_materials/2006-12-08/ agenda_2_registration_update_nov.pdf.

308

Competitive Electricity Markets

provided through transmission owners. Nearly half of the SCR/ICAP resource is provided through demand-side aggregators. The New York market had a total peak demand of 33 939 MW in 2006. Participants in these two programs responded to ﬁve events in 2006 (Isser, 2006). SCR participants are provided at least 2 hours’ notice and receive the zonal real-time price for the duration of an event or 4 hours, whichever is greater. NYISO can activate the EDRP in response to a shortage of operating reserves or a major emergency, and will strive to provide at least 2 hours’ notice. When a curtailment is requested, EDRP participants receive compensation for a 4 hour period. For 2 hours or the duration of an event (whichever is greater) the participant receives the higher of the zonal real-time price or $500 per MWh. The zonal price is paid for the remainder of the 4 hour period. With a peak load of 144 796 MW, the PJM market is the largest of these three markets, and its reliability-focused programs are also much larger. PJM operates Emergency Load Response (ELR) and Active Load Management (ALM) programs.13 Within the ELR, participants can receive both an energy payment and an ALM credit, or simply receive an energy payment (in the “energy-only option,” which waives non-performance penalties, or “ALM deﬁciency charges”). Participants specify their notice time requirements. The minimum duration of a curtailment is 2 hours. Consumers without interval data recorders (IDRs) are permitted to participate in this program on a pilot basis. Each of these markets in the Northeast offers programs focused on economics as well. Each offers day-ahead and real-time markets, which permit demand-side participation.14 Offers generally require speciﬁcation of the quantity, price, start-up cost payment, and minimum run-time information. In most of these markets, there is a $1000/MWh bid price cap.15 Offers to the real-time market must be submitted 1 hour in advance in NYISO and PJM, and 20 minutes in advance in ISO-NE. ICAP resources must submit bids into energy markets. Similar opportunities exist for loads to participate in energy markets in the Midwest ISO (MISO) market. In California, comprehensive sets of reliability-focused programs have been implemented at the utility level, but with extensive coordination with the ISO. About 1800 MW of demand reduction can be achieved through utility-administered reliability-focused programs.16 Commercial and industrial customers can voluntarily sign up for their utilities’ interruptible rate program and receive a reduced rate in exchange for reducing their energy demand to a ﬁrm service level under emergency conditions. In most cases, there are limits on how many times or how many hours they can be called on within a speciﬁed period of time. Currently, there are about 900 MW signed up statewide with the utilities. The second type of program is aimed at residential and smaller commercial customers that will allow their air-conditioner to be turned off for a speciﬁed number of minutes per hour under emergency conditions. They are typically compensated based on a ﬂat payment. There are about 900 MW in those programs as well, aggregated statewide. The sheer coverage of some of the demand response initiatives launched by investorowned utilities in California is quite impressive. Table 8.4 provides a summary of the programs offered by Southern California Edison Company.

EBL

13

Figures in this paragraph are from: PJM (2006).Load response activity report, January through September 2006. 30 September, at: http://www.pjm.com/committees/working-groups/dsrwg/ downloads/20061026-item-06-dsr-activity-thru-sept-2006.pdf. Further information is posted at: http://www.pjm.com/committees/working-groups/dsrwg/dsrwg.html. 14 Demand-side participation in the ISO-NE market may not yet be fully implemented. 15 Some of this information is taken from Isser (2006). 16 E-mail correspondence from Greg Fishman of the California ISO to Rebecca Farrell, 1 December 2006.

Table 8.4. Demand response initiatives of Southern California Edison

’83

•

•

’01

•

’01 ’79 ’01

• •

’01

Ind. (>500 kW)

Comm. (>200 kW)

Comm. (<200 kW)

Residential

Yes

EBL • •

Penalty

Advance notice

Interval metering req.

reduction

Minimum load •

• • •

• • •

Utility service

•

Market

Direct access

•

events

Limited number of

Pay for performance

’87

Eligibility Agricultural and pumping

Agricultural and Pumping Interruptible Air conditioner Cycling Program – Base Air conditioner Cycling Program – Enhanced Base Interruptible Program Large Power Interruptible Optional Binding Mandatory Curtailment Scheduled Load Reduction Program Demand Bidding Program California Power Authority Demand Reserves Program Critical Peak Pricing Critical Peak Pricing – Fixed or Variable SCE Energy $mart Thermostatsm

Guaranteed

Year

Program

payment/ discount

Features

•

•

•

No

•

•

•

•

•

•

No

•

•

•

•

•

•

• • •

• • •

Yes Yes Yes

•

•

• • •

•

•

•

•

•

No

•

•

•

•

• •

•

• •

• •

• •

No Yes

•

• •

• •

• •

• •

Yes Yes

• •

•

•

’03 ’03

•

’03 ’03

•

• •

• •

• •

’02

•

•

•

•

•

Yes

• •

•

•

•

• • • •

•

•

Source: Mark Wallenrod, SCE, "SCE’s Demand Response Programs and Resource Planning," PLMA Fall 2003 Conference, NY, Sept. 8, 2003.

310

Competitive Electricity Markets

In MISO, reliability-based programs are also implemented at the utility level. Many ISOs in the United States are beginning to rely on demand-side resources as a source of operating reserves. As discussed in the following section, interruptible loads have provided operating reserves in the ERCOT market for over 5 years. The wholesale market of the Western Electricity Coordinating Council (WECC) allows interruptible loads to be used for non-spinning reserves, and pilot programs are under way to test the feasibility of using interruptible loads as spinning reserves (Koszalka, 2006). In May 2006, the PJM market expanded opportunities for interruptible loads to provide operating reserves (Neenan et al., 2006), and the NYISO has committed to allowing customers to offer some ancillary services soon.17 Further, ISO-NE is implementing a pilot to explore the feasibility of using dispatchable loads as ancillary services.18 In restructured wholesale markets where retail competition has not been introduced, some challenges have to be overcome in coordinating utility tariffs and programs with the ISO’s efforts to preserve reliability. Who has the button? An interesting development in some restructured markets has been the establishment of real-time pricing as the default service for industrial energy consumers (Barbose et al., 2006). This has been successfully introduced in New Jersey, Maryland, Pennsylvania, Delaware, New York, and Illinois (in the service areas of Ameren and Commonwealth Edison; Folk, 2006). There has also been a proposal to use critical peak pricing as a default pricing scheme in California (Messenger, 2006). Greater direct exposure to market prices may serve to increase demand response. Ontario is implementing the “smart meter” infrastructure necessary for future implementation of either simple time-of-use, critical peak pricing, or real-time pricing on a systemwide basis for residential energy consumers (Cook, 2006), as is California. With this infrastructure, a number of innovative pricing approaches could be extended to smaller energy consumers. Recent analyses suggest that despite all the recent policy pronouncements, advances in technology, and other attention afforded demand response, the actual amount of load involved in dispatchable demand response programs focused on providing system reliability has fallen in recent decades (Barrett, 2006). About 5% of customers in the United States and Canada are involved in some type of demand response program, including direct load control, interruptible tariffs, curtailment programs, and time-of-use rates.19 This amount is lower than in the mid-1990s. The US DOE reports that total potential load management capability in the United States has fallen by 32% since 1996, although concerns have been expressed over some of the self-reported data contained in that report.20 It is also insightful to note that the electricity markets that have experienced the greatest declines in (at least dispatchable) demand response capability are the markets that have undergone the most extensive restructuring, such as ERCOT, the Mid-Atlantic Area Council (MAAC), and the Northeast Power Coordinating Council (NPCC). In many of these markets, regulated interruptible tariffs were terminated, energy efﬁciency activities were disrupted, and load management infrastructure became a stranded cost as a result of restructuring. It has proven difﬁcult to rebuild these types of resources to their prerestructuring levels. Thus, restructuring activity may be at least partially responsible for the initial drop in load management capacity. On the other hand, the ISOs, under FERC’s

EBL

17

Ibid. Ibid. 19 FERC (2006), p. viii. 20 US DOE (2006), p. xii. 18

Demand Participation in Restructured Markets

311

policy direction, are helping to rebuild demand response capability. Moreover, the demand response programs administered by ISO may be recognized as a more-solid resource than some of the interruptible or curtailable tariffs that may have been utilized less frequently or contained buy-through provisions.21 Figure 8.1 provides FERC’s estimates of the changes in demand response capability in various markets between 1998 and 2003.22 As the ERCOT case study presented later in this chapter suggests, it is very difﬁcult to rebuild a large base of demand-side resources into a market once the regulated tariffs and programs designed to foster demand response are terminated. Certainly, demand response initiatives are not conﬁned to North America. European and Asian utilities have operated load control, curtailment, and interruptible programs for many years. Some of the more active utilities in this ﬁeld have included electricity providers that have not undergone restructuring in the nations of France,23 Taiwan, South Korea,24 South Africa, and Japan. As utility systems outside of North America have been restructured to introduce competition, similar challenges have been faced, and similar strategies to promote demand response have been tested. In October 2002, the Nordic Council of Ministers launched an Action Plan for Peak Production in the Nordic Electricity Market, with an initial task of promoting price-elastic demand in the spot market (Bergström, 2003). Statnett, the Norwegian system operator, has used interruptible loads to provide operating reserves during high-demand periods since 2000 (Walther and Vognild, 2005). In this weekly market, offers are based on area, price, and size of curtailable load.

EBL 8%

1998

6%

2003

4% 2% WSCC

SPP

SERC

NPCC∗

MAPP

MAIN

MAAC

FRCC

ERCOT

0% ECAR

Percent of peak load

Interruptible and direct load control by NERC regin 10%

Fig. 8.1. Load management in NERC forecasts. Source: FERC (2006).

21

Charles Goldman contributed some of the insights reported here. The acronyms used in Figure 20.1 are: ECAR = East Central Area Reliability Coordination Agreement; ERCOT = Electric Reliability Council of Texas; FRCC = Florida Reliability Coordinating Council; MAAC = Mid-Atlantic Area Council; MAIN = Mid-America Interconnected Network; MAPP = MidContinent Area Power Pool; NPCC = Northeast Power Coordinating Council; SERC = Southeast Reliability Council; SPP = Southwest Power Pool; and WSCC = Western States Coordinating Council. 23 The tarif vert of Electricite de France was probably the ﬁrst time-differentiated electric utility tariff. 24 The Korean Electric Power Company has operated curtailment programs and innovative rate design for many years. 22

312

Competitive Electricity Markets

In Australia, the National Electricity Market Management Company (NEMMCO) relies upon 373 MW of reserves from demand-side resources (Energy Users Association of Australia, 2006). Also, a demand bidding experiment has been launched in the Australian market. Italy hopes to transition over 2000 MW of interruptible load formerly served under regulated tariffs into a day-ahead market, a congestion management market, and reserve/balancing markets (Cervigni, 2003). Tokyo Electric Power Company offers a variety of interruptible (“supply and demand adjustment contracts”) and time-of-use rate options for industrial energy consumers. Thermal energy storage is also promoted through rate design. Residential consumers can also select from a menu of time-of-use rate options.25 In summary, there have been a variety of efforts to utilize demand-side resources such as interruptible or curtailable loads in restructured markets. The survey provided above is by no means exhaustive. But it covers some of the better-known successful initiatives. As noted above, the size of demand-side resources in many competitive markets has not yet matched their pre-restructuring levels. Why is this so? Every market is different, and different reasons may be cited for the decline of the demand-side resource base in different markets. Yet, the experiences of the ERCOT market offer some useful insights. 8.4. A Case Study of the ERCOT Market Demand response activities in Texas’ ERCOT market may be of particular interest. ERCOT is widely viewed as a successful restructuring initiative, and had a considerable base of interruptible loads prior to the introduction of customer choice. About 85% of the electricity needs in the largest electricity-consuming state in the United States are satisﬁed through the ERCOT market. Because ERCOT is an “intra-state” market with limited interconnections to other markets in the United States, there is very little federal regulatory jurisdiction over ERCOT. This electricity market has been restructured over the past decade to introduce greater competition in the wholesale and retail segments of the industry and to relax regulatory oversight. Senate Bill 373, enacted in 1995, required the Public Utility Commission of Texas (PUCT) to establish rules to foster wholesale competition and create an ISO to ensure non-discriminatory transmission access and an equitable interconnection process for new generation capacity. In the summer of 1997, ERCOT became the ﬁrst operating ISO in the United States. Sweeping reforms were introduced by Senate Bill 7 (SB 7) (enacted in 1999), which allowed customers of the investor-owned utilities within ERCOT to choose among various retail electric providers (REPs) for a retail supply of electricity beginning 1 January 2002. SB 7 also provided the ISO with much greater centralized control over the wholesale market and prompted the establishment of formal markets for ancillary services and balancing energy (see Adib and Zarnikau, 2006). Texas is home to a large number of industrial facilities involved in chemical production, petrochemicals, reﬁnery operation, air separation, pulp and paper manufacturing, and steel production, which can withstand short interruptions in their electricity supply with modest economic loss. Traditionally, these facilities were served through interruptible tariffs, which provided an electrical supply to the facility at a lower level of reliability

EBL

25

Sources: Personal correspondence with Takashi Yamanaka and Tokyo Electric Power Company (2005). Overview of load leveling activity. Presentation, May.

Demand Participation in Restructured Markets

313

in return for a discounted price. Consequently, the state has a very large base of industrial facilities, which can reduce or curtail electricity purchases in response to either an instruction from the ISO or in response to a price signal.26 Since restructuring, ERCOT has had a mixed record of promoting demand response. The introduction of interruptible loads into competitive markets for ancillary services has been generally successful. However, efforts to perpetuate or introduce other forms of demand response have met with limited success. Prior to the full-scale restructuring of the ERCOT market in January 2002, ERCOT relied upon roughly 3500 MW of interruptible load, group load curtailment programs, direct load control, and other load management programs to maintain reliability. As the ERCOT market was redesigned between 1999 and 2001 to foster competition, there was a fear that this large demand-side resource could be lost. Indeed, restructuring led to the termination of tariffs in the areas of ERCOT opened to retail competition. New policies required utilities to divest from “competitive energy services” such as load management programs. And there was (and still is) confusion regarding who would assume responsibility for overall resource adequacy. The PUCT ordered ERCOT to “develop new measures and reﬁne existing measures to enable load resources a greater opportunity to participate in the ERCOT market” (PUCT, 2000a). The steps taken and their track record are discussed below.

8.4.1. Using interruptible loads as a responsive reserve ancillary service The restructured ERCOT market has probably been as successful as any market in the integration of interruptible loads into markets for ancillary services. Many loads served under interruptible tariffs prior to restructuring are now providing ancillary services to the market. The design of ERCOT’s wholesale market permits Loads acting as Resources or “LaaRs” to compete “head-to-head” against generation resources to provide ancillary services, such as responsive reserves (provided by interruptible loads with under-frequency relays and which also agree to manual deployment or curtailment within 10 minutes of notice) and non-spinning reserves (which can be interrupted by the ISO with 30 minutes of notice).27 LaaRs selected to provide these services through ERCOT’s formal day-ahead market for ancillary services receive the market-clearing price for capacity. Alternatively, LaaRs may be self-arranged by a load-serving entity, in which case they would receive a negotiated price. One LaaR (associated with a facility which also has generation) began providing regulation in late 2006 under a “Controllable Load” pilot program. Since March 2007, LaaRs have been permitted to also provide replacement capacity. Demand-side resources can and have been proposed as substitutes for reliability-must-run generating units, but no proposals have been accepted to date. Demand-side resources are not presently used to manage transmission congestion. In light of the pre-restructuring levels of industrial load served under instantaneous interruptible tariffs and armed with under-frequency relays, it was recognized that as much

EBL

26

Further background is provided in Zarnikau (2006). LaaRs are permitted to provide Regulation on a pilot basis. In theory, LaaRs can also provide replacement capacity, although the systems necessary to permit interruptible loads to provide these services have not yet been fully implemented. 27

314

Competitive Electricity Markets

as 2000 MW of load might be interested and capable of providing responsive reserves. As ancillary services markets were designed, this raised two concerns among the ISO’s system operators and ERCOT’s Reliability and Operations Subcommittee: •

If too large a share of responsive reserve requirements were provided by LaaRs, then there might not be adequate generation resources providing responsive reserves. Generating units with governors are better able to stabilize frequency in response to small deviations in frequency than LaaRs with their off-or-on, all-or-nothing response. Also, machines with mass are needed to maintain the stability of the network. • There may be a possibility of “over-shoot” situations, where too much interruptible load might trip off at the same time and raise frequency to an unacceptably high level. Consequently, limits were placed on the amounts of responsive reserves provided by LaaRs. Initially, this limit was 25% of ERCOT’s requirements for this ancillary service (i.e., 575 MW each hour, given ERCOT’s normal requirement of 2300 MW). Later this cap was raised to 50% of ERCOT’s need for responsive reserves (normally, 1150 MW per hour), as concerns surrounding over-shoot abated (ERCOT Staff for the Reliability and Operations Subcommittee, 2002). In addition, strict qualiﬁcation criteria were introduced to preclude energy consumers whose load level could not be accurately predicted on a day-ahead basis from providing responsive reserves. Within a couple of years after LaaRs were permitted to participate in wholesale market, the 1150 MW cap was reached. Presently, 94 LaaRs (working with 10 scheduling entities) are qualiﬁed to provide ancillary services for a total capacity of 1826 MW. While there has been a good mix of LaaRs by size, it is noteworthy that about one-half of the total quantity of LaaRs is provided by ﬁve very large industrial loads, as noted in Table 8.5. The excess supply of LaaRs relative to the cap has led to problems. As competition among LaaRs intensiﬁed for their limited share of the market for responsive reserves, many LaaRs began offering their interruption capability at increasingly negative prices in hopes of securing a place among selected resources within the bid stack, and in anticipation that a higher-price generation resource would set the market-clearing price that all selected resources receive. However, concerns emerged over the consequences of an offer price as low as $19 155 per MW actually setting the market-clearing price. In such a case, all of the selected resources would then have to pay the market a substantial price. This credit risk led to the imposition of a temporary ﬂoor price that prohibited negative bids. The current prevailing proposal is to continue to prohibit negative bids by LaaRs until ERCOT adopts a nodal structure in 2009. At that time, separate markets could be established for LaaRs and generators providing responsive reserves. The cap and restrictive qualiﬁcation requirements have also led to situations where interruptible loads have been ready, willing, and able to interrupt in order to balance

EBL

Table 8.5. Categorization of LaaRs by size LaaR capacity range 1 to10 MWs 11 to 50 MWs 51 to 100 MW Greater than 100 MW

Number of LaaRs

Total capacity (MW)

66 20 3 5

283 388 185 970

Source: ERCOT staff, "DSWG LaaR Bidding Update," 22 February, 2006

Demand Participation in Restructured Markets

315

supply and demand during reliability problems (e.g., the rolling blackouts that occurred on 17 April 2006), but could not be deployed since they were not selected by ERCOT to provide an ancillary service at that time. The amount of interruptible load available to be dispatched by a utility prior to restructuring is compared to the amounts of load qualiﬁed as LaaRs in Fig.8.2. In ERCOT, 3200 MW of interruptible load was available prior to restructuring (not counting curtailment programs and residential direct load control). This amount dropped to almost nothing after interruptible tariffs were terminated. The amount of load qualiﬁed as LaaR is presently over 1800 MW. However, the cap on LaaR participation providing responsive reserves of 1150 MW tends to reduce the quantity of demand-side resources that the ISO can rely upon to provide an ancillary service or interrupt to an emergency condition at any given time.28 Given the cap on the amount of responsive reserves that can be provided by LaaRs and limits to the quantities of other ancillary services required by ERCOT, the pre-restructuring levels of interruptible load cannot be attained unless new reliability-based programs are created. The total amount of interruptible load that the ISO can rely upon in an emergency is presently around 1200 MW (the amounts of LaaR load that normally provides responsive reserves and non-spinning reserves). As noted in Fig. 8.1, ERCOT formerly had a relatively high-demand-side resource that could be relied upon in an emergency, compared to other markets. Today, such demand-side resources available to the ISO are quite limited in ERCOT.

EBL

Interruptible load in ERCOT 3500

Megawatts

3000 2500 2000 1500 1000 500

Ye

ar 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05

0

Year Fig. 8.2. Interruptible load in ERCOT over time. Sources: Data for 1985–93 came from PUCT 1996 Statewide Electrical Energy Plan for Texas, June 1996, and represents interruptible loads plus a small contribution from various load cycling programs. Interruptible load data for 1994–99 came from PUCT Project 22209 Annual Update of Generating Electric Utility Data, 2000. PUCT (2000b) Data for 2000 and 2001 came from ERCOT Capacity, Demand, and Reserve reports for those years. Data for 2002–06 came from Krein, S. (2006). Load participation in ERCOT ancillary services markets. AESP Brown Bag Seminar, 18 April, 2000b. This ﬁgure also appears in Zarnikau, J. (2006). Using interruptible load as an ancillary service in the restructured ERCOT market. US Energy Association Dialogue, July.

28

Note that a small, but increasing, amount of load provides regulation service and non-spinning reserves.

316

Competitive Electricity Markets

8.4.2. Load response to price signals or programs offered by retail electric providers ERCOT’s market structure provides some incentives for large energy consumers to reduce power purchases during peak or high-price periods, including:29 •

The design of transmission charges, which are based upon the consumer’s contribution to demand during four peak times in summer months. • The ability to purchase balancing energy (through a REP) and the design of the wholesale market settlements system rewards qualiﬁed scheduling entities (QSEs) that can reduce generation needs during high-price periods. • Participation in demand response programs sponsored by REPs. A large industrial energy consumer’s transmission charge is based upon the consumer’s contribution to ERCOT’s coincident peak demand in four summer months. Often, transmission charges are treated as “pass-through” costs in the contracts offered by REPs. Consequently, larger energy consumers may see direct beneﬁts by reducing their consumption during the four summer peaks, which are used to allocate transmission costs to consumers and REPs. During the initial years of the restructured market (e.g., 2002 through 2005) consumers were free to deviate from scheduled load levels (in response to price changes, for example) with minimal penalties. “Passive load response” refers to a customer’s deviation from its scheduled or anticipated load level in response to price signals (e.g., balancing energy prices or peak demand periods used to assign transmission costs) in situations where the customer has not formally offered its response to the market as a “resource.” If the actual load level of a QSE turns out to be lower than its scheduled load level during a given 15 minute interval while its actual generation is equal to its scheduled generation, then the QSE is entitled to a payment or credit based on the energy imbalance multiplied by the balancing energy market price. This may provide energy consumers with an incentive to respond to wholesale market prices, provided their REP agrees to settle the consumer separately from other loads served through the REP.30 This separate settlement normally requires the metering of a load’s consumption at 15 minute intervals. At market-open in 2002, IDRs were required on energy consumers with a billing demand over 1 MW. The IDR threshold was later reduced to 700 kW. Some industrial energy consumers rely on balancing energy (essentially, spot market power) to meet some or all of their electricity needs, actively monitor 15 minute balancing energy prices, and reduce electricity purchases when prices exceed threshold levels. The degree to which this practice is permitted has been subject to changing policies since the start of customer choice.

EBL

29

In addition to the programs and activities mentioned here, there are a number of small demand response programs in ERCOT, which have nothing to do with the competitive market. These include Austin Energy’s direct load control program, interruptible tariffs and programs offered by “nonopt-in entities,” which include most of the municipal and rural electric cooperative systems in the ERCOT region, and the Load Management Standard Offer Programs which are regulated energy efﬁciency programs offered by transmission and distribution utilities including TXU Electric Delivery and AEP-Texas Central. 30 Loads are permitted to submit formal offers into the market for balancing energy to reduce their consumption if their offer is accepted. However, there is no incentive for loads to participate in this formal market, since they can achieve the same economic beneﬁts through passive load response without exposure to potential penalties for failing to comply with market rules in the formal balancing energy market.

Demand Participation in Restructured Markets

317

During the ﬁrst years of the restructured market, there was a “balanced schedule requirement” (although some load-serving entities ignored it). Later, a “relaxed balanced schedule requirement” was introduced, in part to encourage REPs and large loads to rely in part on balancing energy to provide near-real-time price signals and foster demand response. Under the relaxed balanced schedule policy, a load-serving entity could elect to purchase a share of its generation requirements from the balancing energy market. While this leaves the REP un-hedged and exposed to price ﬂuctuations in the balancing energy market, many REPs and large energy consumers found this strategy advantageous, particularly if they served loads with some capability to reduce energy usage in the face of high market prices. While volatile, balancing energy prices tend to be lower than the average cost of ﬁrm generation obtained from bilateral contracts. However, following the April 2006 blackouts, ERCOT adopted a more risk-averse operating strategy through which it intentionally biased its short-term load forecast upward and greatly increased its reliance upon replacement capacity in increasing operating reserves. The practice of assigning the cost of procuring replacement capacity to QSEs who were “short” of scheduled capacity at the time the replacement capacity is needed greatly increased the cost of relying upon balancing energy to meet generation needs. True “demand bidding” is permitted, whereby consumers can submit a formal offer to the balancing energy market describing a strike price at which they would curtail and a curtailment amount.31 Under the balancing up load (BUL) program, a load that submits an offer and is struck can also receive a capacity payment based on the prevailing price of non-spinning reserves. However, the capacity payment has not been sufﬁcient to induce consumers into submitting formal offers. Rules requiring all BULs under a QSE’s control to be scheduled as a group and complicated baseline formulas (modeled after those developed for the New York ISO’s Emergency Curtailment Program) have also been cited as impediments to program participation (Trefny, 2006). Instead, voluntary or passive load response is generally preferred. Finally, some industrial consumers of energy participate in curtailment programs that are established by REPs. These are conducted outside of the formal ERCOT market and are used by REPs to shape their generation needs and reduce their costs. If an energy consumer opts to offer its interruption capability into an ancillary services market, then its ability to react to wholesale balancing energy prices, avoid the four summer peaks, and participate in any REP-sponsored demand response programs will be constrained. If the load is providing responsive reserves, then ERCOT monitors the load’s level every 3 seconds to ensure that the load is available for interruption should the system need to rely upon the interruption to maintain frequency. A QSE could incur a penalty (a scheduling control error) if it is not providing its committed level of operating reserves. Thus many of ERCOT’s most ﬂexible, interruptible, or potentially price-elastic electric loads will not react to prices. This appears to be most true for LaaRs, which have a bilateral contract to provide responsive reserves to a REP/QSE and are self-scheduled. Two-thirds of the LaaRs providing ancillary services are self-arranged by a REP or QSE through bilateral contracts in this manner. The amount of load that is actively responding to price signals is thought to be relatively small. A quantiﬁcation of the price elasticity of demand of the 20 largest industrial energy

EBL

31

Note that a formal day-ahead market will not be established in ERCOT until 2009.

318

Competitive Electricity Markets

consumers in Houston to wholesale electricity prices found that one or two are clearly responding to wholesale prices (Zarnikau et al., 2007). Price elasticities have also been estimated for the aggregated block of all energy consumers in ERCOT with IDRs (representing roughly one-half of the demand for electricity in ERCOT). The average own-price elasticity for the aggregated block of all energy consumers in ERCOT with IDRs is quite small (Zarnikau and Hallett, 2007). Based on some simple comparisons of the aggregate load levels of transmission voltage (large industrial) energy consumers between days of likely 4 CP charges and adjacent days, the ERCOT staff has identiﬁed about 600 MW of aggregate demand response, or about a 1% reduction in demand (Jones et al., 2006). The deviations of large loads from their forecasted or scheduled levels in response to wholesale balancing energy prices, to avoid the four summer peaks, and participate in any REP-sponsored demand response programs have caused some scheduling and operating problems. The ERCOT staff has been unable to factor these responses into its short-term load forecasts32 (although it is likely that bad weather forecasts have a much greater contribution to ERCOT’s forecasting error than consumer response to price signals). These changes in demand in response to price changes complicate the ISO’s task of matching supply and demand in real-time. 8.4.3. Demand-side resources that failed to make the transition A variety of programs and tariffs that the regulated vertically integrated utilities relied upon for peak clipping or energy shifting failed to survive in the restructured market. The direct load control program operated by Houston Lighting and Power Company (HL&P) prior to restructuring was sold to Comverge (pursuant to the Commission’s Competitive Energy Services rules). However, Comverge was unable to continue the program due to difﬁculties inherent in dealing with the many REPs who providing power to the program participants, extensive measurement and veriﬁcation requirements placed upon the program by the ERCOT stakeholders, and problems in securing payments from the REPs and other market participants who were likely to realize a beneﬁt from interruptions. Similarly, curtailment programs were terminated as those utility organizations were unbundled. Energy efﬁciency programs and tariffs designed to promote the installation and operation of thermal energy storage were terminated at the start of restructuring, and diurnal differentials in balancing energy prices have been insufﬁcient to interest REPs in providing programs or pricing incentives to promote such technologies in the competitive market.

EBL

8.4.4. New programs under study for today’s zonal market A conﬂuence of factors sparked renewed interest in new demand response program opportunities in 2006. The PUCT voted to pursue an energy-only resource adequacy approach and higher caps on wholesale prices, which places greater reliance on demand response.33 During rolling blackouts on 17 April 2006, there were industrial loads that were ready, willing, and able to be interrupted to restore reliability, but ERCOT had no mechanism to interrupt them. Many interruptible loads were not providing an ancillary 32 33

ERCOT Staff report to the PUCT, Open Meeting of 23 August 2006. See Chapter 9 by Adib, Schubert, and Oren (this volume).

Demand Participation in Restructured Markets

319

service at the time, and there was no program through which curtailments could be requested. Actual planning reserves dipped below 3% during the summer of 2006. And, as noted earlier, a large “waiting list” of industrial interruptible loads that were willing and able to participate in the crowded ancillary services markets had developed. In April 2007, ERCOT, with the support of the PUCT and despite opposition from many stakeholders, sought to implement an Emergency Interruptible Load Program (EILP). However, the offers to provide the service (about 150 MW of demand reduction) fell below ERCOT’s minimum program participation level of 500 MW and the program was not operated during the months of April and May 2007. Additional solicitations for program participants will be held later in 2007. A Tiered Frequency Response (TFR) program has been proposed by the Steel Mill Coalition. Under the proposed TFR program, interruptible loads would set their underfrequency relays to higher settings than loads providing responsive reserves and would commit to long-term contracts.34 Program participants in the EILP would curtail within 30 minutes’ notice if the second level of a system emergency was declared.35 8.4.5. Demand response in the future nodal market When the ERCOT market transitions to a nodal design in 2009, it is likely to take “one step forward and two steps back.” The new day-ahead energy market may open up opportunities for demand response by industrial loads with predictable and ﬂexible energy needs. However, opportunities for loads to respond to real-time prices will become limited. Testifying on behalf of the staff of the PUCT, the Commission Consultant, David Patton, explained the concern over permitting demand response in real-time:

EBL

Passive demand response occurs when loads reduce their consumption in response to prices they observe without actively submitting price-sensitive offers into the wholesale market. The prices that are posted are based on the load levels that ERCOT observes and the generation and load resource offers it has received from suppliers. If ERCOT posts a price 10 minutes in advance and loads respond, then the price for that interval is incorrect. For example, imagine on a high load day, ERCOT posts a price of $200 per MWh 10 minutes in advance. Further, assume that observing this price, load reduces its consumption by 2000 MW. Finally, assume that the price would have been $100 per MWh had the 2000 MW not been included in the dispatch. In this case, dispatch signals will cause generators to over-generate by 2000 MW. In addition, ERCOT will not be revenue-neutral. Generators will have received payments for 2000 MW of generation that will not be collected from the loads, since loads are not consuming that amount. Hence, an uplift charge will be needed to collect the additional revenue needed to compensate the generation.36 For these reasons, the Texas Commission has been unwilling to permit demand response in real-time as it reﬁnes its wholesale market structure, even though the Commission has 34

PUCT Docket No. 32615: Petition of Chaparral Steel Co., Structural Metals, Inc., and Nucor Steel for the Adoption of a Rule to Implement a Tiered Frequency Response Service. 35 PUCT Docket No. 32853: Evaluation of Demand Response Programs in the Competitive Electric Market. 36 Rebuttal testimony of David Patton on behalf of the Staff of the Public Utility Commission of Texas, PUC Docket No. 31540: Proceeding to Consider Protocols to Implement a Nodal Market in ERCOT Pursuant to PUC Substantive Rule 25.501. November 2005.

320

Competitive Electricity Markets

endorsed demand response as a policy goal and ERCOT has explicitly endorsed demand response in real-time as a feature of the future market structure. To reduce forecasting error and to prevent loads from responding to prices in real-time, advanced notice of prices will be eliminated. Also, a penalty (the Reliability Unit Commitment Capacity Short Charge) will be used to discourage REPs (and their customers) from relying upon the market (as opposed to bilateral contracts and the forthcoming day-ahead market) to secure generation. Further complications arise from the use of zonal prices to settle energy purchases, while nodal prices are used to establish the value of a supply-side resource in the market. While taking these steps to discourage price-chasing may provide system operators with better demand forecasts, any demand response in real-time is likely to be sacriﬁced. As the PUCT approved “nodal protocols” that included the features mentioned above to discourage demand response, it nonetheless expressed interest in exploring new avenues for demand response. Through a new project, the PUCT is considering mechanisms to: •

Insulate loads that are deemed to be price-sensitive from the Reliability Unit Commitment Short Charge. • Provide some advance notice of real-time prices to loads (outside of the day-ahead market). • Consider a proposal consistent with priority pricing, whereby loads would provide the ISO with a commitment to curtail at certain price points, in return for protection from various penalties and other incentives (Zarnikau, 2005a).

EBL

8.4.6. Opportunities for improvement ERCOT has not yet succeeded in regaining its pre-restructuring levels of demand response. ERCOT went from having one of the nation’s largest demand-side base prior to restructuring to the smallest overall “existing demand response resource contribution” of any market in the United States, at about 3% of peak demand(see FERC, 2006, p. 87). The need for vibrant demand response in ERCOT is only increasing. The PUCT is pursuing an “energy-only” resource adequacy mechanism, which places considerable emphasis on demand response to balance supply and demand over the long run. In addition, ERCOT is probably the nation’s most concentrated market, and present problems with supplier market power are likely to become exasperated under a nodal market structure. Power plants under common ownership tend to be geographically clustered, reﬂecting the old service areas of the traditional utility suppliers. There is concern among some observers that a wholesale pricing scheme that better recognizes transmission constraints will also tend to divide the ERCOT market into smaller sub-markets, where market share among suppliers could be more concentrated. Clearly, a lot of work remains to ensure that North America’s most successful restructured market provides adequate levels of demand-side participation. Yet, some optimism may be in order. New programs to facilitate demand-side participation are being explored. The PUCT appears willing to reconsider some of the policy decisions that may be responsible for inhibiting demand response. Some of the larger transmission and distribution utilities in the areas of ERCOT opened to retail competition plan to deploy advanced metering systems. TXU Electric Delivery hopes to develop the nation’s “ﬁrst automated,

Demand Participation in Restructured Markets

321

smart electric grid.”37 The new Center for the Commercialization of Electric Technologies is working to develop and capture the beneﬁts of advancing technologies in electric energy transmission, distribution and end-uses. New digital meters under study could be used to remotely control air-conditioning settings or activate “smart” appliances. 8.5. Conclusions Restructuring has been pursued in part based on the premise that by unleashing the creative forces of competition, new pricing and service options will be provided to energy consumers that better reﬂect market economics. Exposure to market prices will foster economically efﬁcient consumption decisions. Market forces, including a demand curve more accurately reﬂecting true price elasticity of demand rather than short-term inelastic demand, will balance supply and demand in operations and planning. Competition can be enhanced through demand-side participation in energy markets. In markets that have restructured and introduced competition at the retail level, these goals have been achieved to a limited degree to date. New pricing options and features are indeed available in markets where customer choice has been introduced. Retailers realize that certain load-shaping actions can reduce their generation costs, as well as the cost of serving their retail customers. ISOs have introduced new curtailment programs to improve system reliability. New competitive markets for ancillary services and installed capacity credits permit loads to compete against generation assets to provide resources to the system. There are now plenty of examples of beneﬁcial demand response efforts in restructured markets. Yet, while policymakers around the world recognize the importance of fostering demand response in competitive markets, the actual implementation of the market structure, rules, programs, technology, and policies necessary to achieve greater demand response has been extraordinarily slow and difﬁcult. While there has likely been a general decline in dispatchable demand-side resources in North America over the past decade, this decline is most pronounced in the markets that have achieved the greatest degree of restructuring. Although the presence of bilateral arrangements between retailers and energy consumers “outside of the market” makes it difﬁcult to accurately quantify the amount of demand response occurring in the competitive retail markets, it appears as though the post-restructuring levels of demand response in restructured markets is somewhat lower than the pre-restructuring amounts achieved through regulated interruptible tariffs, realtime pricing tariffs, time-of-use tariffs, thermal energy storage promotional programs, and load management demand-side management programs. On a positive note, the potential for greater demand-side participation in electricity markets is greater than ever. Widespread deployment of smart metering systems will enable demand response by smaller energy consumers. Technology will advance. Hopefully, our understanding of consumer behavior and the policies and market structure necessary to enable demand response will similarly advance. Many of the new demand response programs administered by ISOs have demonstrated the value of curtailments or interruptions as a resource with considerable value, dispelling some traditional concerns that interruptible loads won’t really interrupt when the system needs them to do so. As restructured

EBL

37

TXU Electric Delivery to Procure 400 000 Automated BPL Meters. Transmission and Distribution World, October.

322

Competitive Electricity Markets

markets mature and the potential value of demand response is better appreciated (or, if generation adequacy mechanisms continue to falter), perhaps we shall see an appreciable expansion in participation in demand response. REFERENCES Adib, P., and Hurlbut, D., (this volume). Market power and market monitoring. Chapter 7. Adib, P., Schubert, E., and Oren, S. (This volume). Resource adequacy: Alternative perspectives and divergent paths. Chapter 9. Adib, P. and Zarnikau, J. (2006). Texas: The most robust electricity market in North America. In Electricity Market Reform: An International Perspective (P. Sioshansi, ed.). Elsevier. Baughman, M., Siddiqi, S., and Zarnikau, J. (1996). Advanced pricing in electrical systems. IEEE Transactions on Power Systems 12(1), 489–502. Barbose, G., Goldman, C., and Neenan, B. (2004). Survey of utility experience with real time pricing. LBNL-54238, November. Barbose, G., Goldman, C., and Neenan, B. (2006). The role of demand response in default service pricing. The Elec. J., 19(3), 64–74, April. Barrett, L. (2006). Load response resources still lagging. Barrett Consulting Associates Report, July. Bergström, M. (2003). DR in competitive electricity markets Sweden, presentation at PLMA/IEA Symposium on Demand Response, September, NY. Brattle Group. (2007). Quantifying demand response beneﬁts in PJM. Prepared for PJM Interconnection and the Mid-Atlantic Distributed Resources Initiative, 29 January. Caves, D., Eakin, K., and Faruqui, A. (2000). Mitigating price spikes in wholesale markets through market-based pricing in retail markets. The Elec. J., 13(3), 13–23, April. Cervigni, G. (2003). Demand response in the design of the Italian electricity market. Presentation at Workshop on Enhancing Demand Response in Liberalised Electricity Markets, Paris, February. Cook, G. (2006). Ontario, Canada’s plan to implement time-of-use rates. AESP Strategies Newsletter, July. Energy Users Association of Australia (2006). Energy users are helping keep the lights on the summer. Press release, January, available at: http://www.euaa.com.au/press_releases.htm. ERCOT Staff for the Reliability and Operations Subcommittee (2002). Utilizing high-set load shedding schemes to provide responsive reserve services. November. ERCOT Staff (2006). DSWG LaaR bidding update, 22 February. Energy Policy Act of 2005, Section 1252(f). E Source (1995). Real-time pricing and electric utility industry restructuring: Is the future out of control? Faruqui, A. and George, S.S. (2002). The value of dynamic pricing. The Elec. J., 15(6), 45–55, July. Federal Energy Regulatory Commission (FERC) (2002). Standardized transmission service and wholesale electric market design. Working Paper, 15 March. Federal Energy Regulatory Commission (FERC) (2006). Assessment of demand response and advanced metering, August. Faruqui, A., Chao, H-P., Niemeyer, et al. (2001). Analyzing California’s power crisis. Energ. J., 22(4), 29–52. Folk, S. (2006). Illinois General Assembly authorizes residential real-time electricity pricing. ASEP Strategies Newsletter, May. Hirst, E. (2002). Barriers to price-responsive demand in wholesale electricity markets. Edison Electric Institute, June. ISO-New England (2006a). Controlling Electricity Costs. Staff White Paper, 1 June. ISO New England (2006b). ISO New England/NEPOOL demand response working group meeting. Presentation by the Demand Response Department, 6 December. ISO-NE, FERC Electric Tariff No. 3, Section III, Market Rule 1 – Standard Market Design, p. 263, available at: http://www.iso-ne.org/regulatory/tariff/sect_3/section_iii_mr1_standard_ mkt_design.pdf. Isser, S. (2006). Demand response survey. Draft presentation, 15 December.

EBL

Demand Participation in Restructured Markets

323

Jones, S., Wattles, P., and Krein, S. (2006). ERCOT emergency load response. PUCT Demand Response Workshop, 15 September See: http://www.puc.state.tx.us/electric/projects/32853/ 091506/ERCOT.pdf Koszalka, M. (2006). Load control as reserves in the west. AESP Brown Bag Seminar presentation, 18 April. Krein, S. (2006). Load participation in ERCOT ancillary services markets. AESP Brown Bag Seminar, 18 April. NARUC (2000). Resolution regarding equal consideration of demand and supply responses in electricity markets. July. Neenan, B., Boisvert, R., and Cappers, P. (2002). What makes a customer price responsive? The Elec. J., 15(3), 52–59, April. Neenan, B., Cappers, P., and Anderson, J. (2006). Demand response in ancillary services markets. AESP Brown Bag Seminar presentation, 18 April. Neumann, S., Sioshansi, F., Vojdani, A., and Yee, G. (2006). How to get more response from demand response. The Elec. J., 19(8), 24–31. NYISO (2006). November 2006 demand response registration. Price Responsive Load Working Group presentation, 8 December, available at: http://www.nyiso.com/public/webdocs/ committees/bic_prlwg/meeting_materials/2006-12-08/agenda_2_registration_update _nov.pdf. Oren, S., Smith, S., and Wilson, R. (1986). Priority service: Unbundling quality attributes of electric power. EPRI Report EA-4871, November. O’Sheasy, M. (2003). Demand response: Not just rhetoric, it can truly be the silver bullet. The Elec. J., 16(10), 48–60, December. Parviz, A., Jaussaud, D., and Hurlbut, D (this volume-a). Market power and market monitoring Chapter 7. Parviz, A., Schubert, E., and Oren, S. (this volume-b). Resource adequacy: Alternative perspectives and divergent paths Chapter 9. Patton, D. (2005). Rebuttal testimony on behalf of the Staff of the Public Utility Commission of Texas, PUC Docket No. 31540: Proceeding to consider protocols to implement a nodal market in ERCOT pursuant to PUC Subst. R. 25.501, November. PJM. (2006). Load Response Activity Report, January through September 2006. 30 September, available at: http://www.pjm.com/committees/working-groups/dsrwg/downloads/20061026-item06-dsr-activity-thru-sept-2006.pdf. Public Utility Commission of Texas (1996). 1996 Statewide electrical energy plan for Texas, June. Public Utility Commission of Texas (2000a). Final order in Docket No. 23220: Petition of the Electric Reliability Council of Texas for approval of the ERCOT protocols. Public Utility Commission of Texas (2000b). Project No. 22209, Annual update of generating electric utility data, PUCT (2000b). Public Utility Commission of Texas (2006). Docket No. 32615: Petition of Chaparral Steel Co., Structural Metals, Inc., and Nucor Steel for the adoption of a rule to implement a tiered frequency response service. Public Utility Commission of Texas (2007). Docket No. 32853: Evaluation of demand response programs in the competitive electric market. Rosenzweig, M., Fraser, H., Falk, J., and Voll, S. (2003). Market power and demand responsiveness: Letting customers protect themselves. The Elec. J., 16(4), 11–23, May. Schweppe, F.C., Caramanis, M.C., Tabors, R.D., and Bohn, R.E. (1988). Spot Pricing of Electricity. Norwell, MA: Kluwer Academic Publishers. Spine, H. N. (2002). Demand response: The future ain’t what it used to be – or is it? The Elec. J., 15(6), 78–86, July. Tokyo Electric Power Company. (2005). Overview of load leveling activity. Presentation, May. Trefny, F. (2006). BUL program status. Presentation at PUCT Demand Response Workshop, 8 December, available at: http://www.puc.state.tx.us/electric/projects/32853/32853.cfm. US Department of Energy. (2006). Beneﬁts of demand response in electricity markets and recommendations for achieving them: A report to the U.S. Congress pursuant to Section 1252 of the Energy Policy Act of 2005. February. Wallenrod, M. (2003). “SCE’s Demand Response Programs and Resource Planning,” PLMA Fall 2003 Conference, NY, 8 September.

EBL

324

Competitive Electricity Markets

Walther, B. and Vognild, I.H. (2005). Statnett’s option market for fast operating reserves. Demand Response Dispatcher, IEA DSM Task XIII Project, April. Zarnikau, J. (2005a). Long-Term BUL offers. Presentation to the ERCOT Demand Side Working Group, October. Zarnikau, J. (2005b). Testimony in Docket No. 31540: Proceeding to consider protocols to implement a nodal market in ERCOT pursuant to PUC Subst. R. 25.501. November. Zarnikau, J. (2006). Using interruptible load as an ancillary service in the restructured ERCOT market. US Energy Association Dialogue, July. Available at: http://www.usaee.org/pdf/Aug06.pdf#13d. Zarnikau, J., Baughman, M., and Mentrup, G. (1990). Spot market pricing of electricity. Forum for Applied Research and Public Policy, 5(4), 5–11. Zarnikau, J. and Hallett, I. (2007). Aggregate customer response to wholesale prices in the restructured ERCOT market. Draft presentation, January. Zarnikau, J., Landreth, G., Hallett, I., and Kumbhakar, S. (2007). Industrial energy consumer response to wholesale prices in the restructured Texas electricity market. Energ. – the Int. J.

EBL

Part III Capacity, Resource Adequacy and Investment

EBL

EBL

This page intentionally left blank

Chapter 9 Resource Adequacy Alternate Perspectives and Divergent Paths PARVIZ ADIB,1 ERIC SCHUBERT,2 AND SHMUEL OREN3 Automated Power Exchange, Santa Clara, California, USA; 2 BP Energy Company, Houston, Texas, USA; 3 University of California, Berkeley, USA

1

Summary

EBL

Ensuring that deregulated wholesale markets meet resource adequacy targets has been a concern of policymakers for more than a decade. This chapter reviews the difﬁcult and controversial evolution of capacity resource adequacy mechanisms in the United States. Next, the alternative used in Texas is examined: an energy-only resource adequacy mechanism based on the successful Australian model. The necessary conditions for a sustainable energy-only approach and a potential transition mechanism to an energy-only approach are also discussed. 9.1. Introduction While electric industry restructuring has been increasingly reﬁned across the world for about 15 years, there is still a signiﬁcant debate on whether direct capacity remuneration should be an integral part of competitive wholesale electricity markets. This chapter will address this question. The authors present a selected historical account of resource adequacy mechanisms throughout the world with a focus on developments in the United States, where the capacity framework has been going through substantial evolution and policy debate in multiple markets under the Federal Energy Regulatory Commission (FERC) jurisdiction for almost 10 years. The chapter also examines the reconﬁrmation of an energy-only framework adopted in Electric Reliability Council of Texas (ERCOT), in part because of the concern that the evolving capacity framework in other US markets was not consistent with ERCOT’s deregulated wholesale and retail markets. In the economic literature, capacity markets are often viewed as a ﬁx to the “missing money” problem where generators are not making enough revenue to recover their 327

328

Competitive Electricity Markets

investment or attract new needed investment.1 There is no evidence, however, that capacity mechanisms actually promote investment in new capacity. Furthermore, some critics claim that such markets impose an extra and unnecessary cost on customers with no proven beneﬁts. Moreover, Moran and Skinner in Chapter 11 present a strong case that the Australian electricity market has enjoyed ample private investment in new generation for the past decade using an “energy-only” resource adequacy mechanism. The newly proposed capacity mechanisms at the Pennsylvania-New Jersey Maryland (PJM) Interconnection and the Independent System Operator in New England (ISO-NE) extend the lead time and duration time of procuring additional resources. While these capacity mechanisms go a long way toward encouraging participation by new entrants, nonetheless, they still reinforce the old central planning paradigm that relies on multiyear constrained optimization models rather than on competitive energy markets to meet resource adequacy needs. This chapter shows that the slow and incomplete transformation of wholesale electricity markets have limited the options that policymakers have had available in creating sustainable resource adequacy mechanisms in the United States. Capacity mechanisms in the United States have evolved in part to overcome shortcomings in transmission construction, lack of adequate demand response programs, and policies that restrict the ability of new generation to interconnect with the transmission grid. The authors contend that these shortcomings are a direct consequence of the fragmented oversight of electricity markets in the United States. In markets that have completed that transformation into a commodities market, ding Australia, Alberta, ERCOT, and New Zealand, an energy-only mechanism has been implemented and appears to be functioning with few problems. In an energy-only resource adequacy mechanism, inframarginal energy revenues and scarcity rents derived from occasional high spot market prices are relied upon to address resource adequacy concerns. This chapter also addresses what many policymakers consider the most contentious and troublesome issues of a sustainable energy-only approach: distinguishing between scarcity pricing and high prices caused by the exercise of market power, increasing marketbased demand response, and developing a backstop mechanism that limits the transfer of wealth from load to generators in years when the reserve margin is tight. In light of the California market meltdown in 2000–01, state and federal policymakers have been reluctant to consider allowing market participants to make energy offers into real-time energy markets that are high enough to allow an energy-only resource adequacy mechanism to be successful. The chapter ends with a proposal to use a bilateral contracting requirement for load serving entities (LSEs) that includes an energy market-based capacity product (“call option”) as a backstop mechanism to transition to a sustainable energy-only resource adequacy mechanism. The proposal requires, however, that the underlying wholesale electricity market complete the transformation into a commodities market that could support such an approach. This chapter is divided into ﬁve sections. Section 9.2 discusses concerns regarding resource adequacy. Section 9.3 presents the evolution of installed capacity markets in the US. Section 9.4 proposes market alternatives to capacity mechanisms. Section 9.5 discusses

EBL

1

The term “missing money” was popularized by Dr. Roy Shanker. See Roy J. Shanker (2003) Comments of on Standard Market Design, Resource Adequacy Requirement. FERC Docket No. RM01-12-000, January 10. Also see Cramton and Stoft (2006).

Resource Adequacy: Alternate Perspectives and Divergent Paths

329

energy-only resource adequacy mechanisms with a focus on the experience in ERCOT. Section 9.6 presents a potential transitional mechanism based on energy-only markets with contracting obligations. Section 9.7 presents conclusions. 9.2. Concerns Regarding Resource Adequacy Reliability of electric service has been a priority of the regulators since the inception of electric industry regulation. The regulatory approach evolved in an era when the pace of technological innovation was slower than today, and when economies of scale in the industry supported a structure of natural monopolies with vertically integrated utilities and captive load, as described in Chapter 1. In the face of rapidly changing technology and increasing opportunities for choosing wholesale suppliers – or in some cases, competitive retailers – there has been movement across the globe to unbundle, either structurally or functionally, regulated utilities. Furthermore, through the introduction of formal wholesale and retail electricity markets, market forces were allowed to determine the mix and quantity of resources to serve enduse customers. An unintended consequence has been the potential for LSEs to lean too much on the spot markets to meet their electricity demand rather than procure sufﬁcient resources through bilateral contracts. Insufﬁcient bilateral contracting by LSEs can lead to a cumulative, system-wide shortfall of resources during times of peak electricity use. Within this paradigm, regulators need to address the maintenance of a prudent reserve margin that meets the reliability needs of newly restructured markets while avoiding imposition of unnecessary expenses on customers and without excessively interfering with customer choice. In the absence of a mandatory reserve requirement to serve end-use customers, regulators need to provide incentives to ensure that sufﬁcient and appropriate mix of resources are developed in the right locations in the grid to maintain the reliability of the grid at a competitive cost over time. The following subsections describe historical approaches under regulatory regimes, identify the problem of “missing money” and lack of adequate incentives to attract investment, highlight perspectives on resource adequacy mechanisms, and describe the prevailing top-down resource adequacy mechanisms adopted in the United States and other countries.

EBL

9.2.1. Historical approaches under regulatory regimes Prior to the introduction of wholesale and retail electricity markets, regulators required utilities to maintain a target reserve margin,2 ranging somewhere between 15% and 25%, to ensure a desired level of reliability and resource adequacy. Regulators ensured that integrated utilities would serve retail load and meet prevailing reliability standards at just and reasonable prices. This approach required a mix of resources designed to serve the ﬂuctuating demand at least cost. New capacity expansion plans were based on forecasted load growth and reserve margin requirements so as to reliably meet future needs. Regulators set the price of electricity service so as to provide the utility with an opportunity 2 Reserve margin and capacity margin are measures used in the electricity industry to determine the percent of additional capacity above expected demand. Reserve margin measures excess capacity as a percent of expected demand while capacity margin shows excess capacity as a percent of existing capacity.

330

Competitive Electricity Markets

for cost recovery and reasonable return on its investment.3 In some jurisdictions, up to the middle of the 1990s, utilities could build new generating facilities only if the new additions were approved in their Integrated Resource Plans (IRPs), which required utilities to review alternatives to constructing generating facilities including purchases from non-utility qualifying facilities, demand-side management, renewable resources, etc.4 Sending timely and accurate price signals to the majority of customers was not a key priority. However, in maintaining a healthy reserve margin, regulators used concerns over reliability and cost as justiﬁcation for their policy of special tariffs for interruptible load, which were primarily targeted at large commercial and industrial customers. The regulatory approach was appropriate in an era when regulators were primarily concerned with controlling monopoly utility proﬁts, technological innovation was slow, and economies of scale in generation and transmission justiﬁed viewing the electric power industry as a natural monopoly. However, technological innovation leading to improvements in the efﬁciency and cost of relatively small power plants and the promise of further technological innovation in metering control, generation, and transmission technologies prompted market reforms that allow LSEs to take advantage of the wider mix of resources available to meet their future needs. The complexity of choices, including price options, now feasible at reasonable cost, can reﬂect the diversity of consumer tastes, preferences, and the willingness of producers and consumers of electricity to accept more risk in exchange for proper rewards. Unfortunately, such expanded choices in wholesale and retail markets open the door for free riding by LSEs when it comes to risk management and public goods such as service reliability. The situation is potentially more complicated in most of the deregulated retail electricity markets, where regulators may not require that competitive retailers procure resources that provide deliverable energy to their loads in a multi-year time frame.5 The glut of generation in some of these newly restructured markets may have also reinforced the practice of relying on spot market procurements and avoiding the burden of long-term contracts in serving load that might switch suppliers in the future. Within this paradigm, regulators need to maintain a prudent reserve margin that meets the reliability needs of newly restructured markets while not imposing unnecessary expenses on customers that may also unnecessarily interfere with customer choice. At present, most regulators do not enforce a reliability requirement on LSEs similar to what the integrated utilities in the old regulated world had to meet. In the absence of a mandatory reserve requirement to serve end-use customers and to maximize proﬁts by reducing unnecessary expenses or increasing market share, some LSEs may have strong incentives to procure a convenient mix of resources, rather than a mix of resources that will allow a wholesale market to maintain reliable service at reasonable cost. This desire to maximize short-term proﬁts also may lead to underinvestment in “iron

EBL

3

An exception to this uniformity was an interruptible tariff for industrial load. For instance, the Public Utility Commission of Texas as well as many other state regulators required regulated utilities to rely on competitive solicitation to procure additional resources to meet their increasing growth in demand for electricity. Competitive solicitations would include conventional generation resources, new technologies, such as renewable resources, and demand-side management programs. 5 When retail markets are still relatively young, competitive retailers are uncertain about their market shares and might be at a competitive disadvantage if they procure ﬁrm resources 2 or 3 years in the future while their competitors might not do so and while loads have the ﬁnancial incentive to buy from retailers with the lowest prices. 4

Resource Adequacy: Alternate Perspectives and Divergent Paths

331

on the ground” at certain times in the electricity business cycle. In the absence of any mandatory requirement to maintain a healthy reserve margin, some LSEs have the ability to lean heavily on the spot market to the extent that their credit limits permit rather than securing resources through long-term contracts.6 Furthermore, since widespread forward contracting has the effect of suppressing real-time prices, such excess reliance on the spot market may be quite tempting. Such free ridership arises from the fact that current market rules and the installed metering and control technologies do not allow the exclusion of those who do not pay for reserves from enjoying the beneﬁts of reserves.7 9.2.2. Problem of missing money Most wholesale electricity markets have some form of price or offer caps and other forms of market mitigation measures to address price spikes or potential market failure resulting from excessive concentration of generation ownership and inelastic demand for electricity (Table 9.1). Unfortunately, mitigation designed to protect the public against market power abuse will also often suppress legitimate scarcity rents and inframarginal proﬁts. In a competitive unrestricted market, scarcity rent is the difference between the value to consumers of the most valuable MWh that cannot be supplied due to limited generation capacity (i.e., the marginal demand-side offer accepted) and the marginal cost of the most expensive MWh served. In a long-run equilibrium, if we allow generators to collect such scarcity rents by letting demand-side bids set the clearing prices in times of shortage, then the scarcity rents should be exactly what are required to cover the amortized ﬁxed cost of the marginal generating unit. Furthermore, if the technology mix is optimal, i.e., least total cost, then the combination of scarcity rents8 and inframarginal proﬁts,9 which amount to the difference between the market clearing prices and the respective marginal costs, will exactly cover the amortized ﬁxed cost of each generation technology in the capacity mix (not just the peaking units) in the long run (Oren, 2005a). From a commercial point of view, scarcity rents and inframarginal proﬁts are the margins needed above short-run marginal costs that allow

EBL

•

owners of existing generation resources to recover their ﬁxed costs of refurbishing existing plants to keep them operating and provide a fair market return on their capital (i.e., the market equivalent of regulated rate of return); • potential developers of new resources to recover their cost of investment. 6

Other reasons, including rate of new entry and rapidly shifting market share in the newly restructured retail market, low price or offer caps to mitigate prices, signiﬁcant ﬂuctuations in fuel prices, as well as uncertainty over future wholesale market design, also contribute to retailers’ decisions not chosen to enter into many long-term bilateral contracts. 7 The US, Australia, Canada, and other countries are currently seeing the spread of advanced meters capable of providing interval metering data, but this technology will not be fully deployed for a number of years. 8 In a competitive unrestricted market, scarcity rent is the difference between the value to consumers of the most expensive MW that cannot be supplied due to limited generation capacity (i.e., the marginal demand side offer accepted) and the marginal cost of the most expensive MW served. 9 Generators that are not on the margin and whose capacity cost is typically lower than that of peaking units on the margin recover their amortized ﬁxed capacity cost from inframarginal proﬁts (i.e., the difference between the MCPE and their marginal cost) plus the scarcity rents. This payment above marginal cost will both insure generation capacity adequacy and an optimal technology mix of generation.

332

Competitive Electricity Markets

Table 9.1. Price or offer caps in various electricity markets Market

Price or offer cap

Alberta Australia California ISO (CAISO) ERCOT (2006) ERCOT (2007) ERCOT (2008) ERCOT (2009) France ISO-New England Italy Japan Midwest ISO New York ISO (NYISO) Netherlands New Zealand NordPool Pennsylvania–New Jersey–Maryland (PJM) Ontario Philippines Singapore South America (Argentina, Brazil, Chile, and Colombia) South Korea Southwest Power Pool (SPP) Spain United Kingdom

$C1000 $AUS 10 000 $400 $1000 $1500 $2250 $3000 No Cap $1000 No Cap No Cap $1000 $1000 No Cap No Cap No Cap $1000 $C2000 62 000 Pesos $SGD 4500 No Cap No Cap $1000 No Cap No Cap

EBL

Notes: 1. The Canadian markets (Alberta and Ontario) offer caps are in Canadian dollars. A Canadian dollar is worth $US 1.05–$US 1.10. 2. An Australian dollar is worth $US 0.90–$US 0.95 at current exchange rate. The Australian market does limit the amount of money a resource can capture on a weekly basis, after which the $AUS 10 000 offer cap drops to $AUS 100. 3. ERCOT market also has a low offer cap of $500 or 50 times Houston Ship Channel gas price index from the previous day, whichever is higher. See note under Table 9.2. 4. Wholesale electricity market in France does not have any price cap. However, it has regulated retail electricity prices which are kept fairly low by the government. As the wholesale market prices are rising, it becomes very difﬁcult for alternative suppliers to compete with the incumbents. 5. The New Zealand market has no cap, but the highest price allowed in settlement has been $NZ 10 000. At the time of this writing, NZ regulators are reviewing the need for a formal cap. 6. NordPool has no price cap in the ﬁnancial forward electricity market; however, there is a system limitation to the bids (currently EUR 2000) in the physical day-ahead auction. This limit is not viewed as a regulatory price cap, as it can be changed from day to day. Furthermore, NordPool has cross-border transmission capacity constraints between the NordPool countries and on the interface to Germany with market splitting when these cross-border interfaces are congested. 7. 45–50 Philippine pesos equals $US 1.00 at current exchange rates. 8. One $SGD is worth about $US 0.69–$US 0.70 at current exchange rates. 9. At the time of this writing, MISO has ﬁled a plan with FERC that keeps the $1000 offer cap but would raise real-time prices as high as $3500 when MISO is using reserves to clear the energy market.

Resource Adequacy: Alternate Perspectives and Divergent Paths

333

In other words, scarcity rents represent the market mechanism needed to signal resource shortages and provide incentives for new investment in resources.10 Furthermore, this mechanism allows the loads, through demand-side offers, to determine how much generation capacity is needed. Proponents of scarcity pricing argue that in the absence of sufﬁcient demand response, prices in shortage situations should be allowed to reﬂect the Value of Lost Load (VoLL) which could be in the thousands of dollars, well above any existing offer cap currently in place in most restructured electricity markets.11 Mitigated energy prices that often suppress scarcity rents may be insufﬁcient for generation resources to earn enough return to cover their ﬁxed costs, a problem that has been characterized as the “missing money” problem. Other forms of market interference motivated by reliability concerns, such as out of market procurement of resources (e.g., reliability must run or RMR), deployment of operating reserves to avoid involuntary curtailments, and reliability unit commitment (RUC), will have the effect of suppressing spot energy prices and contribute to the “missing money” problem. Consequently, a number of market participants and economists believe that market designs with existing price mitigation measures (that cap prices at $1000 per MWh or less) do not allow sufﬁciently high energy prices to provide adequate cost recovery to existing generators and may not induce timely construction of new resources for new resources.12 For a more detailed discussion of the “missing money” problem, see Cramton and Stoft (2005, 2006). While in principle a liquid forward market could address the “missing money” problem by providing adequate forward support for resource adequacy, most existing restructured electricity markets show limited amount of market-based long-term contracting. Evidently, the protection provided to LSEs by the capped spot prices reduces, from their risk management perspective, the optimal quantity of forward contracts in their portfolios and caps the price they are willing to pay for such contracts. A higher offer cap would transfer more risk to LSEs and would provide incentives for them to cover a larger portion of the quantity risk through bilateral contracts and reduce their reliance on the spot market.

EBL

9.2.3. Contrasting perspectives on resource adequacy mechanisms There are two prevailing approaches to resource adequacy. One is a “top-down” or accounting-based approach that is driven by the question of whether generators are making sufﬁcient income from energy and ancillary services to recover their ﬁxed costs. This perspective, adopted in a number of markets in the US and abroad, directly addresses the “missing money” problem, as elaborated by Cramton and Stoft,13 with minimal changes to the energy market. The top-down approach has opted to solve the “missing money” problem resulting from artiﬁcial suppression of energy spot prices by supplementing generators’ income through “capacity payments” or through the introduction of an artiﬁcial short-term capacity product for which the demand is set administratively and the cost allocated to the load. The income from such a product is intended to make up the missing money and thus ensure the ﬁnancial integrity of generators and attract new investment if needed. 10

For a more detailed discussion on the role of scarcity rents in resource adequacy, see Oren (2005). For example, Hogan (2006) estimates VoLL at $10 000 per MWh. 12 See Joskow (2005); Cramton and Stoft (2005, 2006); Alberta Department of Energy (2005). pp. 26–36. 13 See Cramton and Stoft (2005, 2006). 11

334

Competitive Electricity Markets

Capacity markets are viewed as a ﬁx to the missing money problem but there is no evidence that they actually • •

promote investment in new capacity; do not impose an extra and unnecessary cost on customers.14

The new proposed forward capacity market (FCM) mechanisms at ISO-NE and the reliability pricing model (RPM) proposed at PJM extend the lead time and duration time of the product. In Chapter 10, Bowring discusses the RPM in detail. These reforms go a long way toward encouraging participation by new entrants, but at the same time have moved closer to the traditional centralized resource planning paradigm by being prescriptive with regard to location and even fuel mix of new generation resources. The call option features embedded in the FCM proposal also improve the economic basis of the non-performance penalties and the market reciprocity, whereby generators receiving capacity payments must forgo peak energy rents (PER). However, these mechanisms still have the general shortcomings of capacity markets which suppress incentives for demand response, selfprovision of sufﬁcient generation resources to meet peak load, and tend to hinder risk hedging by load or through competing retailers. Capacity mechanisms have been controversial in the markets where they are being implemented. Critics have stated that they overly compensate existing inefﬁcient generation, rely on prices that are administratively determined, put too much investment risk on LSEs, and have stiﬂed innovation and bilateral contracting for customers.15 For a more sympathetic perspective on capacity mechanisms, see Bowring’s discussion in Chapter 10. The second approach, used in a number of markets including ERCOT, Australia, New Zealand, and Alberta, takes the view that providing resource adequacy is primarily a challenge of managing risk in a competitive commodities market. This “bottom-up” approach starts with the theoretical construct, discussed earlier; implying that in an ideal “energyonly” market, inframarginal proﬁts and scarcity rents will cover capacity costs and lead to an optimal mix of generation technologies. Hence, earnings from unmitigated spot market transactions and long-term bilateral contracts that provide a risk-sharing mechanism between consumers and producers should stabilize generators’ income streams and induce the proper level of investment. Thus, resource adequacy can be ensured by eliminating spot market distortions and by facilitating bilateral contracting and risk management. To the extent that

EBL

• • •

unmitigated energy markets are infeasible, markets for risk are not fully functional, or contracting practices fail to provide adequate incentives for generation investment,

then various forms of contracting and hedging obligations can be imposed. In addition, centralized procurement mechanisms for such hedges can be instituted to promote resource adequacy. However, since such mechanisms rely on energy as the underlying commodity, they can only succeed if artiﬁcial barriers to efﬁcient energy spot prices that can reﬂect scarcity are eliminated. In other words, customer protection through price mitigation measures must be greatly restricted and replaced, if necessary, with 14

See, e.g., PennFuture (25 October 2006) for criticisms of RPM. See ELCON (2006) for comments on PJM, ISO-NE, and NYISO capacity mechanisms from the perspective of industrial customers and PennFuture (2006) for comments focused on PJM’s RPM from the perspective of environmentalists. 15

Resource Adequacy: Alternate Perspectives and Divergent Paths

335

mandatory hedging requirements that will complement voluntary risk management and contracting practices. Unmitigated energy prices will ensure that the capacity payments to generators through voluntarily contracts and mandatory hedges resolve the “missing money” problem. Variants of the latter approach have been implemented in several restructured wholesale electricity markets with no capacity payments but a high offer cap with limited mitigation of resource-speciﬁc offers (i.e., an energy-only resource adequacy mechanism). The experience suggests that high offer caps can indeed provide a strong incentive for generation resources to supply electricity service, and for market-based demand response to be available. There is evidence that high offer caps have resulted in increased voluntary bilateral contracting between buyers and sellers and demand response, which has resulted in lower average spot market prices. Apparently, LSEs faced with the prospect of paying thousands of dollars per MWh on spot market procurement during shortage periods have been compelled to maintain forward supply contracts in order to avoid such outcomes. For instance, a number of retailers in the Australian market hedge their positions with option contracts on peaking generation that require the peakers to make energy offers into the real-time market on behalf of the retailer.16 Similarly, in the ERCOT market, a retailer purchased a multi-year contract for the output of some peaking units to avoid being exposed to the spot market during summer peak.17 9.2.4. Current top-down resource adequacy mechanisms As shown in Table 9.2, there is no uniﬁed approach to resource adequacy in existing wholesale electricity markets in the world. The authors ﬁnd two basic approaches to resource adequacy that attempted to address the “missing money” issue differing by whether they specify quantities or prices for generation capacity.18 Several restructured wholesale electricity markets in the US imposed a relatively low energy offer cap in the late 1990s but introduced a distinct capacity product (i.e., a capacity-and-energy resource adequacy mechanism). Payments for capacity aim to stabilize the payment stream for resource owners and energy prices for consumers, which in turn will attract investment in new capacity. In the northeastern US markets such payment for capacity is implemented indirectly through a capacity obligation imposed on the LSEs and an adjunct market for trading and procurement of capacity credits. Recognizing that consumers are only interested in energy consumption while capacity is an artiﬁcial product, capacity markets are based on an administrative demand prescription according to a technical determination of target capacity. The administrative demand can be set at a ﬁxed level as in the traditional ICAP markets and the current ISO-NE FCM proposal. But recently, several ISOs have adopted a demand function approach, also known as variable resource requirement (VRR), where the administrative demand for capacity is price-sensitive, thus allowing more capacity to be purchased as the procurement price declines.19

EBL

16

Authors’ communication with Peter Adams, Manager, Surveillance and Enforcement, Markets Branch, Australian Energy Regulator, 1 February, 2005. 17 The retailer’s name is not divulged for reasons of conﬁdentiality. 18 A comprehensive analysis of alternative resource adequacy mechanisms with particular focus on Europe can be found in the Ph.D. thesis of De Vries, L.J. (2004). Correljé and De Vries discuss this issue in Chapter 2. 19 Such an approach is used by the NYISO and is part of the PJM Reliability Pricing Model (RPM).

Competitive Electricity Markets

336

Table 9.2. Some examples of restructured electricity markets with and without capacity payments Level of offer/Price cap In some cases, cost-based Energy offers Low offer caps (up to $1000) Medium offer caps ($1000 to $3000) No or high offer caps (above $3000)

No price mitigation

Capacity market

Direct capacity payments

No capacity payments

South America, Spain, Italy, South Korea Alberta, ERCOT (2006)

ISO-NE, NYISO, PJM

Ontario, Philippines, Singapore Australia, France, Japan, New Zealand, Netherlands, NordPool

Ontario

Alberta, CAISO, ERCOT (2006), MISO, SPP ERCOT (2009), Philippines, Singapore Australia, France, Japan, New Zealand, Netherlands, NordPool, United Kingdom

Notes: 1. South American markets include Argentina, Brazil, Chile, and Colombia. 2. The Australian market does limit the amount of money a resource can capture on a weekly basis, after which the $AUS 10 000 offer cap drops to $AUS 100. 3. In CAISO, Load Serving Entities (LSEs) are subject to a forward bilateral contracting obligation. 4. The ERCOT market does limit the amount of money a resource can capture to $175 000 per MW of capacity, after which the $3 000 offer cap drops to low offer cap of $500 or 50 times Houston Ship Channel gas price index from the previous day, whichever is higher. 5. At the time of this writing, MISO has ﬁled a plan with FERC that keeps the $1000 offer cap but would raise real-time prices as high as $3500 when MISO is using reserves to clear the energy market. 6. In South Korea, capacity payment is determined annually by the Generation Cost Evaluation Committee considering the long-term marginal ﬁxed cost of the generators, which varies by fuel type.

EBL

The alternative capacity remuneration mechanism is direct capacity payment originating from the theory of socially optimal peak load pricing in a regulated monopoly dating back to Boiteaux (1960). Under that theory, social welfare is maximized when energy is priced at marginal cost. However, since such pricing cannot recover ﬁxed costs, a secondbest approach with minimum social welfare degradation is to supplement marginal cost pricing with a capacity payment that equals to the amortized ﬁxed cost of the marginal technology and recover these added costs as a demand charge from consumption during peak periods. Capacity payments are popular in several South American countries such as Argentina, Chile, Peru, and Colombia, in European countries such as Spain and Italy, and in South Korea where generators receive direct payments for capacity from the system operator, which are uplifted to customers on a prorated basis. Capacity payments are sometimes differentiated according to generation technologies (e.g., South Korea has a two-track payment for base-load capacity and peaking capacity) and may be commingled with stranded cost recovery (e.g., Spain). Often such payments are coupled with a cost-based offer requirement in the energy market (e.g., South Korea and Peru).

Resource Adequacy: Alternate Perspectives and Divergent Paths

337

In Argentina and Peru, only capacity that is selected to produce energy or provide reserves will receive capacity payments. This practice has resulted in Argentina of energy offers below marginal cost. Such “underbidding” is prohibited in Peru where the regulator determines the marginal cost of each generation technology except for suppliers with natural gas plants, who can declare their marginal cost annually. In South Korea there is a two-tier system for base-load and peaking generators. All generators are required to submit regulated marginal cost-based offers for energy. Generators are classiﬁed into two tiers, which receive different capacity payments, and their energy offers are cleared separately for base-load and for peaking load, resulting in two different market clearing prices for the generators. The load is charged an average price that recovers the energy and capacity payments to the generators. In general, capacity payment mechanisms are an effective means to stabilize the income of incumbent generators, but there is no clear evidence that such payments encourage investment in new generation. Generators receiving capacity payments have no obligation to use that income for new investment or for improvement in their facilities. Furthermore, since capacity payments have the effect of suppressing marginal cost and resulting energy spot prices to consumers who cannot opt out of paying the capacity charges, such mechanisms undermine potential demand response. In Chile, one of the ﬁrst countries to liberalize their electricity industry and to institutionalize a capacity payment scheme, recognition that this scheme has not produced the desired outcomes in terms of providing incentives for investment has led to recent re-assessment of their resource adequacy policy and new proposals along the line of call option obligations.20

EBL

9.3. Evolution of Installed Capacity Markets in the US In the early days of deregulated wholesale markets in the eastern US, capacity owners were under a must-offer obligation to offer their capacity into the market for a given price ($/MW). Such arrangements evolved from resource sharing agreements that existed in long-standing power pools such as PJM and NEPOOL, now ISO-NE. In these agreements utilities had an obligation to maintain sufﬁcient capacity to cover their peak loads and were penalized for capacity shortfalls. The establishment of a market for short- term capacity credits, later referred to as an Installed Capacity Market, or ICAP, was intended to enable utilities with excess capacity to be compensated by utilities with capacity shortfalls when the overall capacity available in the system was sufﬁcient to meet total peak load and reserves requirements. Such market, however, had several shortcomings. Because of their short duration, only existing “steel in the ground” could participate, resulting in the so-called “bipolar pricing,” where the price was either zero when the total available capacity exceeded the total requirement or as high as the shortfall penalty when there was a capacity shortage in the system. In addition, while the total capacity might have met the aggregate capacity requirement in the system, transmission constraints often made available capacity non-deliverable to load pockets where it was needed, resulting in local reliability problems. Because of the short duration of ICAP products, ICAP markets were also susceptible to exercise of market power due to non-contestability by new entrants. At the same time, energy markets have not been able to provide correct price signals and sufﬁcient revenue to provide incentives for new entrants due to the lack of scarcity pricing 20

See Barroso et al. (2006).

Competitive Electricity Markets

338

that would reﬂect deployment of operating reserves to prevent energy shortfalls. These deﬁciencies have been widely recognized and, over the last several years, signiﬁcant efforts have been devoted to improving the capacity markets in the US, as described later. In this section, the authors ﬁrst review earlier versions of ICAP payments established in the northeastern US markets, particularly in PJM. This review will also include minor modiﬁcations that were made to these mechanisms to address some of the earlier problems. Next, we will brieﬂy discuss new and improved version of ICAP markets that were developed through further co-operation between stakeholders and state and federal regulators to improve the performance of these capacity market mechanisms. The mechanisms discussed later include the NYISO’s extended product duration and sloped demand function, the PJM’s 3 year forward-looking capacity market (Reliability Pricing Model or RPM), and the ISO-NE’s 3 year forward-looking capacity market (FCM). The following subsections describe the early ICAP markets and their minor reﬁnements, followed by a discussion of recent steps taken in some markets to introduce new and improved capacity mechanisms. 9.3.1. Early ICAP markets and their reﬁnements The PJM installed capacity market was the ﬁrst of its type to become operational in 1999. Capacity owners were under obligation to offer their capacity into the market for a given price ($/MW). There was no distinction between old and new capacity, and capacity owners would be penalized if they could not meet certain performance standards. The ICAP payments did not take into account the value of the location of the resource or the operational characteristics of the unit. For a more detailed discussion of capacity markets, see Bowring’s discussion on capacity markets in Chapter 10. This approach proved inadequate for assuring resource adequacy, as the US Federal Energy Regulatory Commission (FERC) noted in its 2004 Staff report, stating that

EBL

Much of the country has no obvious market mechanism to signal the need for new building in advance of shortages. The success of capacity markets in addressing the issue is not yet proven.21 9.3.1.1. Shortcomings of the original PJM ICAP The original PJM ICAP market had the following four shortcomings: 1. 2. 3. 4.

Bipolar nature Deliverability problems Lack of contestability Lack of scarcity pricing

Bipolar nature of ICAP clearing prices The ﬁrst incarnation of an ICAP market has been the most contentious aspect of ICAP. This version of ICAP was a monthly product. From the generators’ viewpoint, a one-month installed capacity obligation does not provide sufﬁcient certainty of cash ﬂow to attract capital for new generation investment that would be built for years in the future. Furthermore, the short look-ahead period of the ICAP obligations in the early design of the PJM market left no room for participation in the 21

Federal Energy Regulatory Commission (2005). p. 65.

Resource Adequacy: Alternate Perspectives and Divergent Paths

339

ICAP market by the newly planned capacity due to the length of time that it takes for such capacity to be built. Consequently, the ICAP markets have been limited to existing capacity, and the result of that was bipolar pricing: the value of the ICAP product was either zero (during times when the reserve margin was healthy) or inﬁnity (during times when the reserve margin was thin).22 The high price volatility and eventual collapse of the initial daily ICAP market at PJM led to the development of a more sustainable monthly capacity market and to a proposal for a seasonal capacity obligation.23 Similar moves toward capacity products with longer durations have been implemented at ISO-NE and NYISO. The PJM ICAP market currently operates with multiple maturities: 12 months, 6 months, and 1 month, as well as a daily market. The NYISO operates markets for capacity obligations of 6 months, multiple months up to 6 months, and 1 month. Deliverability problems of ICAP products Inability to build transmission quickly enough to accommodate new generation investment and concentration of new capacity in locations that are remote from load pockets where generation capacity was needed results in deliverability problem in meeting expected electricity demand. For example, in ISONE new quick response units were built in Maine near the gas sources but energy from these resources could not be delivered to Connecticut load centers because of transmission constraints. Lack of contestability with new generation As indicated earlier, the short-term nature of the capacity product made it impossible for new generation capacity to participate in the ICAP market so that incumbent generation with steel in the ground could exercise market power in shortage situations. The reforms extending the duration of the ICAP obligation did not go far enough in terms of creating capacity obligations that could enable responses by the newly planned investment when ICAP prices increased due to capacity shortages.

EBL

Lack of scarcity pricing In cases of energy shortage the system operators tend to draw energy from operation reserves. Using reserves augments the balancing energy offers curves and results in suppressing scarcity rents that would have resulted if the shortage was mitigated through demand-side offers. To avoid such suppression of price signals, energy spot prices must reﬂect the cost of reduced reliability resulting from the depletion of operating reserves, but such scarcity pricing was not introduced in the early market designs. Similar price suppression results from deployment of “reliability must run” units by the system operators in order to assure local reliability when markets fail to provide needed local resources. Over time, it has become apparent in US electricity markets that setting clearing prices based on tightly mitigated resource-speciﬁc offers do not send appropriate scarcity pricing signals in real-time when the system operator needs to deploy operating reserves. Nodal markets in the eastern Interconnect of the US are instituting various forms of scarcity pricing to send a more appropriate price signal in real-time, allowing prices to rise more frequently to the $1000 offer cap in each market, improving short-term reliability of the grid 22

While the value of the product would be inﬁnity if the market price were unconstrained, in reality the price of the monthly ICAP product cleared at the cap set by PJM. 23 Federal Energy Regulatory Commission (2002).

340

Competitive Electricity Markets

through market-based mechanisms rather than a command-and-control approach.24 These markets are moving toward or have implemented real-time co-optimization of energy and ancillary services. In addition, one or more of the markets use an administrative pricing mechanism that adjusts real-time energy offers to reﬂect whether the real-time energy offer stack is being or has been depleted.25 Additional cost recovery for generation has been a by-product of this change rather than a driving force behind it. 9.3.1.2. Centralized Resource Adequacy Market (CRAM) The deﬁciencies in existing ICAP markets and seams problems arising from incompatibilities in these markets prompted the commissioning of a study to develop a proposal for a single integrated capacity market for the ISO-NE, the NYISO, and PJM. The NERA (2003) report recognized the need for a long-term forward-looking capacity obligation to ensure adequate investments in generation and proposed a central resource adequacy market (CRAM) for these three markets. The report proposed that LSEs be subject to a capacity obligation to assure that sufﬁcient capacity is in place with sufﬁcient lead time for planning and construction, as suggested by FERC in its Notice of Proposed Rule Making that proposed a Standard Market Design. The ISOs would act as central buyers of capacity and make forward commitments to buy capacity. These commitments would be supported ﬁnancially by uplifted charges to all LSEs during the capacity supply period. According to the CRAM proposal: the ISO would determine the resource need in advance of the planning period, would hold a central procurement through an auction, would pay the auction price to all resources provided during the period and would recover the cost from load during the planning period. The difﬁculties arising from uncertainty with respect to load obligations several years in the future would be eliminated and all LSEs would face a common charge for resource adequacy that would be passed on to consumers and would be competitively neutral at the retail level. Consumers will receive the beneﬁts of adequacy and pay the cost of adequacy.

EBL

A key aspect of the proposed scheme was that The planning horizon must be sufﬁciently long to enable the CRAM to be a deciding factor in the decision to construct. …Only when the pool of competitors is expected to include entrants can market power concerns be adequately addressed. Practically, this means that a three-year planning horizon is the minimum. 26 The proposal recommended that the commitment period should be from 1 to 3 years with preference for longer durations in order to reduce generators’ uncertainty about revenues – which is expected to result in lower risk premiums in their costs of capital. The proposal suggested that all required capacity be procured or under contract at all times, arguing that sequential auctions for progressive procurement would be unreliable for determining 24 McNamara, Ron (2006). Midwest ISO, resource adequacy in Midwest energy markets. Presentation at the Organization of MISO States, May 8–9, 2006. Available at http://misostates.org/R1-% 20Ron%20McNamara%20-%20MISO.pdf 25 Starting in 2005, the NYISO began using an operational reserves demand curve when it is short of operational reserves. ISO-NE, PJM, and MISO have reviewed the issue and are in the process of implementing a similar approach. 26 NERA Economic Consulting (2003). Central Resource Adequacy Markets for PJM, NY-ISO and NE-ISO, Final Report. February.

Resource Adequacy: Alternate Perspectives and Divergent Paths

341

prices. The CRAM proposal was shelved due to disagreements among interested parties in the three North Eastern ISOs. However, some of the elements of the CRAM proposal have resurfaced recently and inﬂuenced capacity market designs adopted by ISO-NE and by PJM. 9.3.1.3. ISO-NE’s proposed locational ICAP market (LICAP) The New England ICAP market has experienced bipolar pricing of its capacity deﬁciency auction like other ICAP markets. Prices were ﬂuctuating between zero and the capacity deﬁciency penalty, but prices eventually collapsed due to system-wide excess capacity, as shown in Fig. 9.1. After February 2004 ICAP prices have essentially dropped to zero. Litvinov et al. (2004) reported that ICAP prices have been insufﬁcient to support existing generation and new investment. Furthermore, because the capacity market in New England does not account for transmission constraints, system-wide excess capacity in the ISO-NE territory has masked local deﬁciency of capacity in congested areas such as Boston. Out of the total generating capacity in New England, which amounts to 32 615 MW, 41% is ﬁnancially distressed due to revenue shortfalls and the owners are in various stages of bankruptcy. Figure 9.2 (reproduced from Litvinov et al., 2004) illustrates the average net revenue shortage for combined cycle (CC) and combustion turbines (CT) units under alternative estimates of the carrying costs for these units. The carrying cost includes cost of capital, plant O&M cost, and other economic parameters such as tax, inﬂation, and risk adjustments.27 The estimates of required ﬁxed cost recovery shown in Fig. 9.2 motivated, in part, the ISO-NE’s proposed solution of creating locational ICAP markets for four distinct areas: Maine, Northeast Massachusetts/Boston (NEMA), Connecticut, and the rest of the ISO footprint. Co-ordinated ICAP auctions account for three directional capacity transfer

EBL

500

ICAP Supply auction results in 2003–2004 ($/MW-month)

400 300 200 100 0 Apr- May- Jun- Jul- Aug- Sep- Oct- Nov- Dec- Jan- Feb- Mar- Apr- May03 03 03 03 03 03 03 03 03 04 04 04 04 04 Obligation month Source: Litvinov, Yang and Zheng (2004)

Fig. 9.1. Prices in the ISO-NE ICAP market.

27

In contrast, according to Potomac Economics(2006). [2005 State of the Market Report for the ERCOT Wholesale Electricity Markets. July, pp. 49–50] wholesale electricity prices in ERCOT in 2005 greatly exceeded the thresholds for ﬁxed cost recovery for base-load coal and nuclear power plants. Given the projections for strong economic growth and natural gas prices in ERCOT for the foreseeable future, market participants have informed ERCOT planning staff that they are considering investing tens of thousands of MWs of new coal-ﬁred, nuclear, and wind capacity in ERCOT through 2010.

Competitive Electricity Markets

342

Net revenue and capacity costs 1999–2003 12.00

$ / kw-month

10.00 8.00

6.88 Unfunded capacity costs

4.48 6.00 5.41

4.00 2.00 0.00

Net revenue from energy

4.01 4.12

4.12

Low case CC

High case CC

1.59

1.59

Low case CT

High case CT

Sorce: Litvinov, Yang and Zheng (2004)

Fig. 9.2. Average net revenue shortage of combined cycle and combustion turbine generators in the ISO-NE.

constraints: Maine exports, NEMA imports, and Connecticut imports. Financial capacity transfer rights (similar to ﬂowgate rights) would be issued to provide instruments for hedging locational ICAP (LICAP) price differences. A demand curve for LICAP was to be used in the ISO-NE spot LICAP auction. According to Cramton and Stoft (2005), the demand curve was intended to provide a rough approximation to a capped annual energy revenue stream, including scarcity rents that would naturally decline as reserve capacity increases and scarcity rents go down. The proposed LICAP demand curve, illustrated in Fig. 9.3, was anchored at a nominal level of target capacity with a corresponding price that equals the amortized expected capacity carrying cost of a peaking unit. From that reference point, currently set to 106% of the minimum LICAP requirement, the curve extends linearly in both directions. The demand curve intersects the $0/kW per month level at some preset capacity value above the target level. Currently, that point is set to 118% of the minimum requirement. The demand curve

EBL

Price 2 × EBCC

EBCC

Target capacity

Capacity

Note: EBCC = expected benchmark carrying cost (annualized fixed cost of frame unit) Source: Cramtonand Stoft (2005)

Fig. 9.3. Proposed ICAP Demand Function in ISO-NE.

Resource Adequacy: Alternate Perspectives and Divergent Paths

343

rises as quantity drops below the target level and is capped at twice the expected carrying cost of a peaking unit when capacity is at or below the minimum LICAP requirement. The LICAP clearing price was to be adjusted ex post on an annual basis by subtracting the inframarginal energy revenues per MW per year realized by a combustion turbine (CT) used as a benchmark. The adjusted clearing price was intended to settle LICAP shortfalls or excesses among LSEs and to determine payments to generators and dispatchable load resources. Payments were to be prorated based on availability during a predetermined set of days when generation capacity is scarce. The locational feature of LICAP and the prorating of payments based on availability on certain days were intended to add intrinsic value to an otherwise artiﬁcial product whose demand is derived from an administrative requirement. Likewise, although artiﬁcially determined, the downward sloping demand function provided an effective means of eliminating the binary nature of ICAP prices, the result from a vertical demand function, and it eliminates much of the incentive to withhold capacity. Eventually the LICAP proposal was scrapped due to strong opposition from different advocacy groups, a strong lobby in the US Congress, and governors in ﬁve out of six states in the ISO-NE jurisdiction. 9.3.2. New and improved capacity mechanisms Given some of the shortcomings mentioned above regarding the earlier ICAP markets in the Northeastern US, market operators and stakeholders along with scholars identiﬁed new approaches to improve the operation of capacity mechanisms. In fact, new and improved versions of previous capacity markets were introduced in all three ISOs (PJM, NYISO, and ISO-NE), and FERC eventually approved their implementations. The improved mechanisms are brieﬂy described in the following subsections.

EBL

9.3.2.1. PJM’s reliability pricing model To improve its capacity mechanism, PJM recently reached a stakeholders consensus in ﬁnalizing its Reliability Pricing Model (RPM) with a variable resource requirement (i.e., sloped demand function), which was ﬁled with FERC on August 31, 2005. RPM is based on an integrated resource planning model that looks to4 years into the future to determine generation resource needs in terms of location and fuel mix. Under the originally proposed scheme, the needed generation capacity to maintain adequate reliability is procured through a central auction on a 4 year forward basis, which enables participation by existing generator and new investors. RPM encompasses existing and planned transmission, generation, and demand-side response, as well as incorporating locational pricing in its forward auction. Bowring discusses this mechanism in more detail in Chapter 10. The RPM auction features an administratively determined downward sloping demand curve that allows the procured quantity to vary with price. This scheme is also known as variable resource requirement (VRR). The use of an administrative demand function has been rationalized on the ground that it reduces volatility of the capacity payment to generators and thus encourages more investment in generation capacity, resulting in increased social welfare.28 However, this argument is debatable since it is based on the assumption that generation ﬁrms are risk-averse, while in calculating the social beneﬁts 28

See testimony by Ben Hobbs in FERC Docket No EL05-148-000 and ER05-1410-000, Initial Order on Reliability Pricing Model.

344

Competitive Electricity Markets

society as a whole and consumers in particular are assumed to be risk-neutral. The intent of RPM was to be compatible with areas having retail choice as well as areas with traditional regulation. After the initial approval by FERC29 on April 20, 2006, the parties entered into 4 months of settlement discussion (ordered by a FERC Administrative Law Judge) that resulted in only slight modiﬁcation to the original proposal. Speciﬁcally, the contracting lead time was reduced from 4 to 3 years and the demand function was shifted down so that it is capped at 1.5 times the cost of new entry (CONE) which is estimated at about $65000/MW per year. The curve drops linearly from 1.5 × CONE at 98% of target capacity to 0.2 × CONE at 105% and then vertically down to zero. The RPM establishes locational capacity requirements, allows for demand response and transmission participation, has explicit market mitigation rules, and allows opt-out alternatives for LSEs that do not want to participate. Features such as seasonal pricing of capacity, operational price adders, and load following requirements for portion of the capacity obligation, which were in the initial proposal were eliminated, while some other minor features have been added. 9.3.2.2. NYISO’s Demand curve model In an attempt to reduce such volatility of ICAP prices that had been at the NYISO and other ISOs, the NYISO was the ﬁrst to introduce a variable resource requirement, also known as an ICAP demand curve, in the New York capacity market.30 The demand curve model was developed through the stakeholder process and came into effect in May 2003. Prior to introducing VRR, the NYISO ran a semi-annual auction for 6 month capacity products and a monthly capacity auction for monthly capacity products for the remainder of the 6 month capability period, as well as a centralized deﬁciency auction prior to each month. Each LSE had to provide contracts to demonstrate to the NYISO that it was covering its capacity requirement for the ensuing month. Any shortfalls were covered through the centralized deﬁciency auction in which the NYISO bid for all the deﬁcient capacity at a price equal to the deﬁciency penalty imposed on LSEs for each MW-month of capacity deﬁciency. LSEs exceeding their capacity obligation could offer their excess in the auction. The deﬁciency auction represented a “vertical demand function” where the ISO demanded a ﬁxed quantity of capacity, and resource providers and LSEs with spare capacity offered supply schedules against it. The experience has been that prices in that auction were either at the deﬁciency price or close to zero. Under the VRR arrangement, the 6 month and monthly ICAP auctions continue to operate as double auctions in which LSEs bid for supply and resource providers offer supply. The deﬁciency auction, however, has been replaced with the VRR, which represents a downward sloping demand curve capped at the deﬁciency penalty. The downward sloping segment of the demand function is linear and determined by two points. The ﬁrst “reference point” is deﬁned by the minimum capacity requirement and a price that equals some percentage of the estimated cost of a CT. The second point is set at the level of capacity at which the capacity value is nil. The parameters of the ICAP demand curve

EBL

29

FERC Docket No EL05-148-000 and ER05-1410-000. New York Department of Public Service (2003a); New York Department of Public Service (2003b); New York Independent System Operator Inc. (NYISO) (2004a); and New York Independent System Operator Inc. (2004b). 30

Resource Adequacy: Alternate Perspectives and Divergent Paths

345

vary by location (speciﬁcally differentiating New York City (NYC) and Long Island from the rest of the state) and are subject to adjustment. According to the NYISO Tariff 2004b (section article 5, section 5.14.1(b)), the reference point for NYC as of May 1, 2004 was set at $151.14/MW per day for 100% of the minimum local ICAP requirement, and from there the curve declined linearly to $0 at 118% of the minimum ICAP requirement. In the NYISO report to FERC31 it is stated that the ICAP demand curve has achieved the goal of stabilizing ICAP spot prices in the deﬁciency auction. Furthermore, purchased quantities in the deﬁciency auction have increased, while clearing prices have decreased. The deﬁciency auction has also provided a price ﬂoor for the 6 month and monthly capacity markets. The VRR seems to function well and mitigates incentives for withholding capacity by rewarding available capacity in excess of the minimum requirement and by recognizing that such extra capacity has value in enhancing reliability and moderating energy and ancillary service prices.

9.3.2.3. ISO-NE’s forward capacity market Likewise, ISO-NE went through a consensus-building process with stakeholders and state regulatory authorities to ﬁnalize a settlement agreement to address its capacity market. In April 2006 a FERC Administrative Law Judge approved this settlement agreement that outlined a forward capacity market for ISO-NE that will replace the previous LICAP design. The new agreement envisions a 3 year forward-looking capacity market (FCM) where the capacity product emulates some key features of physical call options and the procurement is done through a descending clock auction with a vertical demand curve. The FCM has integrated some key features that are similar to the call option approach described earlier and some element of the CRAM approach. Capacity is procured 3 years forward for duration of 1 year for incumbent generators and up to 5 years for new generation through an annual competitive descending clock auction. The procurement is zonal, based on local reliability needs, load forecasts, and forecasted transmission availability and the capacity payment could vary by location. The procured capacity contracts must be backed by qualiﬁed existing or planned physical generation resources or by demand-side resources. The procured capacity contracts entail an obligation to offer energy in the spot energy market at a strike price that reﬂects the marginal energy cost of a generic peaking gas turbine unit. The strike price is enforced by deducting from the capacity payments the peak energy rents (PER), which represents the excess revenues of a generic peaking unit computed for the annual duration of the contract32 . The capacity payments for multi-year contract obtained by new generators are based on the auction price of the ﬁrst year and indexed for inﬂation in subsequent years. Payments for the capacity and cost recovery from the load occur at the time of performance (starting 3 years after the procurement). Periodic and seasonal reconﬁguration double auctions enable parties to adjust their positions by committing additional capacity or withdrawing (delisting) committed capacity. Such delisting can take place up to 4 months prior to start of the commitment period.

EBL

31 FERC (2004). Report on Implementation on ICAP Demand Curve. New York Independent System Operator, Inc., Docket No. ER03-647-000. 32 The PER deduction are capped at the auction-based capacity payment so that the net capacity payment cannot be negative.

346

Competitive Electricity Markets

The seasonal reconﬁguration auction also allows changes in procured capacity so as to reﬂect changes in load forecasts and for trading of obligations among market participants. The traded tender in the reconﬁguration auction is full-year commitment, unlike the traditional ICAP markets (e.g., NYISO) where the traded tender has been monthly or even with shorter-term capacity obligations. Non-performance is subject to penalties in the form of reduced capacity payments. In that respect the FCM contract differs from call options since the non-performance penalties are not based on actual liquidation damages reﬂecting the difference between the spot energy cost and the strike price. The FCM contract is designed to prevent non-performance penalties in excess of the net capacity payment (after PER adjustment) received by the non-performing resource, which would be possible under market-based penalties, reﬂecting actual liquidation damages. A transition mechanisms employing straight capacity payments will be implemented to assure continuity of capacity-based income to incumbent resources. 9.4. Energy Market Alternatives to Capacity Mechanisms The evolution of capacity markets in US East Coast nodal markets has been a gradual realization through experience that undifferentiated capacity and its related energy are insufﬁcient to meet all the requirements of operating a wholesale electricity market efﬁciently and reliably. Furthermore, experiments with retail markets in the eastern Interconnect have been mostly unsuccessful to date, with at least one state, Virginia, moving to re-regulation while others have been in turmoil. The notion that consumer choice will determine the level of reliability and the resulting need for generation capacity seems untenable at this time in many jurisdictions. As a result, regulators have favored centralized planning based on engineering consideration and constrained optimization methods in determining the capacity needs while limiting the role of market mechanisms such as auctions, request for proposals (RFPs), and centralized capacity markets to procurement of these predetermined quantities.33 The evolution of capacity mechanisms to resource procurement mechanisms, such as seen with RPM, raises the question, which we address in the following subsections: “Why not let energy market forces guide resource adequacy decisions?” The following subsections brieﬂy describe energy market-based resource adequacy mechanisms with no direct capacity remuneration and assess the potential beneﬁts of such an approach.

EBL

9.4.1. Energy market-based resource adequacy mechanisms It has been argued that capacity mechanisms will always be a required part of a deregulated electricity market because regulators, policymakers, and the public are reluctant to accept occasional high and volatile energy prices that would be required under an energy-only approach. The concern about high prices in spot markets is clearly a potential political impediment to the implementation of an energy-only resource adequacy mechanism. The 33

Using a multi-year constrained optimization as part of a resource adequacy mechanism is theoretically sound as the duality theory shows that outcomes of a constrained optimization and a set of centralized spot markets are equivalent in the absence of market power. See Charles Rivers and Associates [Ruff, Larry] (2004). A Transitional Non-LMP Market for California: Issues and Recommendations, pp. 5–6.

Resource Adequacy: Alternate Perspectives and Divergent Paths

347

troubled evolution of capacity mechanism in the eastern Interconnect has highlighted, however, a more fundamental issue in meeting resource adequacy in those wholesale markets, which is the halting transformation of electricity deregulation itself. The preference or even necessity for capacity markets that feature a centralized mechanism that relies on constrained optimization over a 3 or 4 year period reﬂects a symptom of basic problems in electricity market infrastructure. After all, one of the drivers of electricity market reforms has been general discontent with the traditional vertically integrated local monopoly-based electric power industry whose decisions and performance incentives were encumbered by the regulatory compact. One of the primary motivations for reforms was the recognition that like in other critical infrastructure industries, market forces could lead to better investment decisions and innovation. Approaches such as RPM and FCM appear to be taking the electricity deregulation away from the fundamental goal of electricity deregulation. The goal of electricity deregulation is the transformation of an industry dominated by regulated monopolies into a competitive electricity market that looks, as much as technically feasible, like a commodities market and can deliver to customers the beneﬁts of economic efﬁciency. A competitive commodities market requires easy entry of new suppliers, good transportation networks, liquid spot and forward markets, and vibrant competition on both the wholesale and retail levels. To function reliably and efﬁciently, a wholesale electricity market needs the right mix of baseload, load-following, and peaking generation. The system operator needs the appropriate mix of generation and load resources to meet real-time contingencies. The following fundamentals, some of which are not part of a number of electricity markets in the US and abroad, are necessary for electricity markets to truly become a fully ﬂedged commodities market that can allow market forces to meet resource adequacy needs while maintaining system reliability:

EBL

•

Easy interconnection of generation: Generation must be able to respond to locational market prices by quickly siting at the appropriate location in the electric grid where additional supply is needed. Difﬁculties in generation siting undermine the power of scarcity pricing signals and can increase the potential for market power abuse. • Proactive investment in new transmission: Delivery of energy to all loads at competitive costs is important to create genuine competition among suppliers for all loads. Many load pockets, particularly in the United States, are associated with areas that are in non-attainment for ozone or other air quality standards, making the siting of new generation in those locations problematic. Proactively reducing transmission bottlenecks and anticipating load growth would encourage the import of competitively priced power from outside of the load pocket. • Socialized payments by all loads of new transmission: In certain markets in the United States, regulatory authorities spend substantial time and effort deciding what portion of a proposed transmission expansion is needed for reliability (the cost of which is paid by loads) and what is needed to improve economic dispatch (the cost of which is paid by generation). This approach greatly reduces the amount and speed of transmission construction and adds risk and cost to the siting of new generation if a generator needs to pay a substantial and undetermined portion of the transmission construction costs. The experience in ERCOT within Texas, Alberta in Canada, and in Australia suggests that socializing the payment of transmission construction allows for smooth and quick entry of new generation resources in the wholesale market. This may turn to be the most economical way to improve market efﬁciency and result in signiﬁcant beneﬁts to consumers by further fostering wholesale market competition.

348

Competitive Electricity Markets

•

Enhancing demand responses: Having price responsive load increases competition for generation at or near peak demand, reducing the need for or scope of ex ante mitigation out of concern for system-wide market power abuse while allowing for scarcity pricing. In Chapter 8, Zarnikau presents a detailed discussion on demand response. • Bilateral forward contracting: Bilateral forward contracting can provide a myriad of energy risk management features customized to the preferences and abilities of end-use customers. Capacity markets that do not use an option approach can greatly restrict the type and size of demand response because of the rigidity of the associated regulatory process and the long lead times for procuring new resources. • Retail competition: Competition removes the agency problem of incumbent utilities that own generation. These utilities may hinder or fail to respond to opportunities to adopt new load management technologies that would undermine the proﬁtability of their generation holdings. Retail competition allows for a wider range of options in pricing the use of electricity by end-use customers by relying on the dynamic creativity of the market rather than a slow and unresponsive regulatory process to meet the preferences of end-use customers. • Regulatory co-ordination: Energy-only markets, such as Alberta, Australia, ERCOT, and New Zealand, have a much stronger degree of regulatory co-ordination than is seen in FERC jurisdictional markets, where the US Federal government is the regulator of the wholesale market and each individual US state government is the regulator of the state’s retail load. Jurisdiction of transmission planning and payment for new transmission is even more fragmented in FERC-jurisdictional markets. Because well-functioning electricity markets are ﬁendishly complex with numerous interrelationships, regulatory co-ordination is vital in developing and maintaining the necessary conditions for an effective energy-only resource adequacy mechanism with retail customer choice.

EBL

9.4.2. Beneﬁts of an energy market-based approach The beneﬁts from using an energy market-based approach to resource adequacy are the efﬁciencies derived from having energy price risk managed by loads and investment risk managed by generation developers.34 As described in the previous subsection, this optimal mix of price and investment risk has failed to materialize in most markets because the market transformation itself has been incomplete. These risks are no greater than those risks associated with other commodity and ﬁnancial markets in the world today provided that the enabling technologies and institutional arrangements for demand side participation in the energy are in place. Managing these risks will encourage market participants to develop more cost-effective ways of delivering and consuming electricity through bilateral contracting while relying on spot markets for adjustments in their position in response to unforeseen load ﬂuctuations or contingencies. The result will be a more efﬁcient use and production of electricity that will serve the same amount of economic activity within a given competitive electricity market at reduced cost. In the regulated world, the rates that customers paid included implicit capacity payments for generating units regardless of whether they were used continuously as base-load units 34

The discussion in this section is based on Schubert (2005).

Resource Adequacy: Alternate Perspectives and Divergent Paths

349

or only a few times a year as peaking units. Although regulators required utility assets to be used and useful, the justiﬁcation for new units presumed an obligation to serve and that all customers, except interruptible customers, were entitled to ﬁrm services. By determining an identical, ﬁxed level of reliability of electricity service for each customer, the regulators implicitly set a value of reliability that was based on statutory and policy decisions and quantiﬁed in terms of engineering criteria. In a deregulated market, units with low-capacity factors need to earn sufﬁcient revenues to be kept online; however, they may not earn enough money because a number of market participants do not wish to pay for a very high level of reliability or are not required to do so by a regulatory rule or market protocol. As an alternative, the market relies more on price-responsive demand and peaking shaving to reduce the number of units with low capacity factors while maintaining system reliability. In reality, if the price of electric service rises to a certain level, some market participants may be willing to curtail their electric service voluntarily for a number of hours each year rather than pay peaking units to deliver that power during those high-priced hours. These market participants would consider the old standard of reliability inappropriate for their needs, for at least a portion of their electric service. An efﬁcient market design would reﬂect the different values of reliability (interruptibility) they place on a portion of their electric service by letting market prices reﬂect those preferences provided that such service quality differentiation is technically feasible. Under an energy-only resource adequacy mechanism, high offer caps could provide needed potential “headroom” for commercial and industrial loads, in the form of priceresponsive loads to compete with the traditional peaking units, the old standby of the regulated era that can sit idle for 90 percent or more of the year. This competition with traditional peaking units would take place in the real-time energy and ancillary services markets in a number of ways: easing ramping constraints in the market, managing load during summer peaks, and providing services to the system operator in the form of reserves and possibly regulation. The possibility of high prices combined with new technologies could provide incentives for smaller loads to increase their deployment of supply- and demand-side alternatives such as solar panels, energy efﬁciency appliances, and controllable loads.35 Higher summertime prices, combined with real-time pricing for a variety of customers, including residential, would provide stronger incentives for installing energy-efﬁcient air-conditioning, increasing insulation of buildings and homes, and creating demand for smart appliances like pumps for swimming pools and hot water heaters that can be set to operate when prices are low.36 Zarnikau presents addition examples in Chapter 8.

EBL

35

An example of a controllable load is an air-conditioning cycling program. When prices and demand rise to a certain level or when market operator declares an emergency condition, a device on an air-conditioner receives a signal, which turns off the air-conditioner for a certain amount of time. When a large number of air-conditioners are aggregated under one controller, the amount of demand responses can be calibrated by the number of air-conditioners cycled at one time. In this capacity, the portfolio of air-conditioners can mimic a small generator by incrementally increasing or decreasing the energy used by the portfolio and could provide ancillary services to grid operators. 36 Time-of-use pricing, with predetermined rates reﬂecting higher average costs of generating electricity over peak summer hours, would be one way for residential load to be exposed to price risk. Large industrial and commercial customers might be in a better position to take the greater risks and reap the greater rewards involved in real-time pricing.

350

Competitive Electricity Markets

9.5. Energy-Only Markets 9.5.1. Existing energy-only markets While various markets have implemented some form of capacity markets, there has not been a clear indication that such capacity markets have actually met their main objective to attract new investment in resources to meet increasing demand in those markets. In contrast, as mentioned earlier, several markets in the US and abroad have been operating without facing serious capacity shortages. Moran and Skinner present a more detailed discussion of the Australian approach to ensuring resource adequacy in Chapter 11. The markets listed in this category share a common feature, namely none of these markets have established formal capacity mechanism to date (i.e., an energy-only resource adequacy mechanism). Market-based energy prices, the sale of capacity services to RTOs, and bilateral contracts for energy are the only avenue for generators in these markets to cover their operating costs, and contribute toward the recovery of their original investment costs. For example, Australia has no capacity payments but a very high offer cap. The offer cap of $AUS 10 000 can provide a strong incentive for resources to provide electricity service and for LSEs to maintain forward supply contracts to avoid paying thousands of dollars per MWh for covering their retail service obligation in the spot market during a shortage. The high offer caps in Australia have increased bilateral contracting between buyers and sellers, which has resulted in lower average spot market prices. In contrast, Alberta has a much lower offer cap of $C1000 with no mitigation of generator offers. New Zealand has no formal offer cap. In Texas, after the passage of legislation in 1999 that deregulated retail competition in 2002 and as ERCOT prepared to operate as a single control area37 , the Public Utility Commission of Texas (PUCT) approved the wholesale market rules with a $1000 offer cap. ERCOT saw a rush on investment in base-load combined cycle capacity without the support of a capacity mechanism. While the reserve margin in ERCOT was sufﬁcient in the years immediately after retail deregulation, concerns were raised that the wholesale market neither sent the appropriate market signals that valued the location (and deliverability) of that investment nor provided sufﬁcient market signals to value the operational characteristics of generation and load resource (e.g., ﬂexibility of real-time dispatch, minimum loading of generation units, and start-up times). In addition, almost all of the new dispatchable generation was designed to operate as base-load with little or no new peaking generation entering the market. In deliberations on the appropriate resource adequacy mechanism to choose for ERCOT, competitive retailers and industrial consumers strongly objected to the use of a capacity market approach. A retailer group suggested even higher offer caps than the PUCT accepted, reﬂecting their strong dislike of a capacity payment approach to resource adequacy. At least one of the PUCT commissioners expressed reluctance to institute a potential subsidy to generation, in the form of capacity payments, which once present would be difﬁcult to remove.38 In 2006, the PUCT, which regulates both the wholesale and retail markets of ERCOT, adopted a combination of market power and resource adequacy rules that explicitly

EBL

37 ERCOT began operating as a single control area on July 31, 2001 when pilot retail competition began in ERCOT. Prior to that date, ERCOT consisted of ten individual control areas operated by each major integrated electric utility in ERCOT. 38 Public Utility Commission of Texas, Project No. 24255, Rulemaking Concerning Planning Reserve Margin Requirements, Memo from Commissioner Barry T. Smitherman, 15 July 2005.

Resource Adequacy: Alternate Perspectives and Divergent Paths

351

rejected capacity payments in favor of raising the system-wide offer cap to ensure resource adequacy.39 In its resource adequacy rule, the PUCT stated that it adapted the Australian energy-only resource adequacy mechanism to the ERCOT market.40 In this proceeding, the Commission adopted Substantive Rules 25.504 (Market Power) and 25.505 (Resource Adequacy). The combination made ERCOT an energy-only market, in contrast to a capacityand-energy approach used in electricity markets in the Eastern Interconnect of the US. In the ERCOT resource adequacy mechanism, the offer cap is to be raised from the $1000 level that prevailed when the rule was adopted in August 2006 to $3000 in 2009. In addition, the Commission ended a system-wide market mitigation measure, the Modiﬁed Competitive Solution Method (MCSM)41 , which changed market clearing prices ex post under certain market conditions that suggested economic or physical withholding might have occurred. The PUCT also expressed its intention to rely on increased market-based demand response to meet its resource adequacy goals.42 Increased market-based demand response also would weaken the potential for market power abuse during times when scarcity pricing was expected. As part of this rulemaking project, the Commission developed a formal deﬁnition of market power, reduced mitigation on smaller market participants, and gave larger market participants the opportunity to apply for the Commission’s approval of voluntary mitigation plans. The rule will raise the offer cap in ERCOT-procured markets to allow generation and load resources the opportunity to recover their ﬁxed costs, improve incentives for bilateral contracting, and increase the transparency of ERCOT-procured ancillary service and energy markets. This is a courageous step by the Texas Commission that is regarded as politically infeasible not only in California, which is still recovering from the 2001 energy crisis, but even in MISO that has recently opted for an energy-only approach.

EBL

9.5.2. Market power abuse and scarcity pricing Because price spikes could be substantially higher under an energy-only resource adequacy mechanism with medium-to-high offer caps, regulators would have an even greater need to distinguish the difference between scarcity pricing and market power abuse. Therefore, there is a need to supplement this ex ante framework with substantial market monitoring resources and greater transparency to proactively address potential market power abuses. Various markets that are using the energy-only approach to resource adequacy have used the following approaches to address potential market power abuses. 9.5.2.1. Transparency of offers into ISO-procured energy and ancillary services markets Increased disclosure will increase market transparency, providing incentives for market participants to make offers into ISO-procured energy and ancillary services markets that are consistent with the properly functioning competitive market and not the result of market power abuse or other market manipulation. The implementation schedule for 39

Public Utility Commission of Texas, Project No. 31972, Order Adopting Amendment to Substantive Rule 25.502, New Substantive Rule 25.504, and New Substantive Rule 25.505, p. 6. A more detailed discussion of the development of this rule can be found in Schubert et al. (2006). 40 Ibid, p. 42. 41 See Hurlbut et al. (2004). 42 Public Utility Commission of Texas, Project No. 31972, Order Adopting Amendment to Substantive Rule 25.502, New Substantive Rule 25.504, and New Substantive Rule 25.505, pp. 68–9.

352

Competitive Electricity Markets

disclosure is also being tied to the schedule for increases to the offer cap, thereby further emphasizing the PUCT’s decision that these two issues are interrelated.43 The interrelationship the PUCT cites is consistent with disclosure policies in electricity markets in the United States and other foreign markets. In FERC jurisdictional markets, for instance, resource-speciﬁc information submitted into an ISO-procured market is released 6 months after the information was gathered, which is consistent with heavily mitigated individual resource offers and a low offer cap.44 Quick disclosure of resource-speciﬁc information appears to provide limited beneﬁt under these circumstances, because market participants are protected ex ante from potential price spikes, know the limited range in which the offers are made, and rely on mechanisms run by an ISO to trigger scarcity pricing in the markets. In contrast, an energy-only resource adequacy mechanism with lighter price mitigation and high system-wide offer caps work best when the ISO discloses resource-speciﬁc offers or unit output quickly. This approach is based on the experience of the Australian market and the belief that companies that have the potential of abusing their market power will be reluctant to expose themselves to public criticism resulting from actions they take in the market to raise prices. This combination of lighter mitigation and quicker disclosure is seen in established electricity markets outside of the United States: the Australian electricity market discloses resource-speciﬁc offers with the names of the generators making the offers within 24 hours; the New Zealand electricity market discloses the same information within 14 days and may shorten the disclosure window in the near future; and the Alberta electricity market displays the output of each generator, by name, on its website in real-time.

EBL

9.5.2.2. Structural components to address market power The Texas Legislature has put in statute a limitation on ownership of no more than 20 percent of installed capacity in the ERCOT market. In addition, the deregulation of the Texas market required integrated utilities that would offer customer choice to unbundle into separate retail, wires, and generation companies with strict code of conduct rules limiting their interaction. 9.5.2.3. Focus market monitoring on key players In ERCOT, the PUCT instituted a provision in its market power rule that stated that market participants that owned or controlled less than 5 percent of total installed generation capacity would be deemed not to have system-wide market power (informally known as “small ﬁsh swim free” provision). As a result, the market monitor can focus more time and energy on the actions of a handful of large players in the market who would be considered most likely to be pivotal suppliers during non-peak system conditions and have the potential of exerting unilateral market power. 9.5.2.4. Voluntary mitigation plans In ERCOT, market participants with portfolios larger than 5 percent of the installed capacity have the opportunity to earn scarcity rents on their units. If they are uncertain whether 43

Public Utility Commission of Texas, Project No. 31972, Order Adopting Amendment to Substantive Rule 25.502, New Substantive Rule 25.504, and New Substantive Rule 25.505, pp. 27–8. 44 Recently, a number of these ISOs have decided to implement scarcity-pricing mechanisms that prescribe speciﬁc situations when prices can rise to the offer cap.

Resource Adequacy: Alternate Perspectives and Divergent Paths

353

the offers from those units would be considered an exercise of market power, they have the opportunity to apply for a voluntary mitigation plan with the PUCT. The PUCT would review the plan and, if approved, would provide the market participant with an absolute defense against charges of market power abuse as long as it adhered to the voluntary mitigation plan.45 9.5.2.5. Keeping market participants informed on short-tem overall supply and demand Market participants are provided short-term forecasts to assist them with their unit commitment decisions (quick-start generation and demand resources). In the Australian market, for instance, the market operator emphasizes the importance of Projected Assessments of System Adequacy (PASA) in informing market participants of unit availability and load forecasts.

9.5.3. Demand participation One of the current problems in restructured electricity markets is the highly inelastic demand for electricity among all but the largest consumers. Zarnikau covers demand elasticity and demand response in Chapter 8. Such inelastic demand requires a substantial generation (as opposed to resource) reserve margin.46 When the generation reserve margin falls, an electricity market is vulnerable not only to high prices (e.g., scarcity pricing) but also to abuse of market power, because differentiating between scarcity pricing and market power abuse is difﬁcult to prove after the fact. Increasing market-based demand response increases competition with generation at near-peak or at peak demand, reducing the need for or scope of ex ante mitigation of potential system-wide market power abuse while allowing for scarcity pricing. An energy market-based resource adequacy mechanism would require widespread and active participation of demand-side resources, which could be encouraged through access to interval metering for residential and small commercial customers through the deployment of advanced “smart” meters. In May 2007, the PUCT adopted an advanced metering rule for ERCOT for this very purpose.47 Such a mechanism increases the rewards for demand-side participation, and as a result allows the market to ﬁnd and use the services of end-use customers with a lower VoLL and a greater willingness to reduce electricity in response to high prices. As a result, the market can avoid involuntary curtailments with a lower system-wide offer cap. Encouraging “reliability at a price” through voluntary, price-sensitive load shedding (curtailable load) would allow market participants to more efﬁciently reﬂect their value of reliability while maintaining the overall reliability of the grid. The amount of curtailable load deployed in a given year would be a function of anticipated prices and would complement the entry and exit of generation resources in a generation building/retirement cycle. Demand-side resources that are available for deployment when prices are high because of a low generation reserve margin could serve as a shock absorber for end-use customers in the face of the time lag of building new power plants.

EBL

45

The Texas Legislature considered legislation that would make this voluntary mitigation provision mandatory for the two largest generation portfolios in ERCOT, however, the proposal failed to receive approval before the end of the Legislative Session. 46 Interruptible customers provided a resource in regulated markets. 47 PUCT Substantive Rule 25.130, Advanced Metering.

354

Competitive Electricity Markets

The thinner the generation margin, the higher market prices would be, which would provide more demand response. As more generation enters the market, it increases the generation margin, lowers market prices, and reduces the beneﬁts for participation of demand resources in the market during the years with abundant capacity. Retailers will compete on managing price risk, with the potential of more customized packages offered to end-users. The result should be a more vibrant retail market. All parties will beneﬁt from less generation capacity with limited use standing by. The same level of economic activity will be served by less generation capacity without sacriﬁcing reliability of customer who places a high value on continuous service. Load management or “peak shaving” programs, which focus on reducing electricity use during summer afternoons, are more effective when prices are predictably high for enough hours to justify an investment in peak-shaving technologies or processes. Zarnikau presents a more detailed discussion of this issue in Chapter 8. 9.5.4. Limiting excessive wealth transfers The time lag between the market price signal and the entry of new generation complicates market mitigation that, if unchecked, could lead to signiﬁcant transfers of wealth to generators. Therefore, the newly restructured wholesale electricity markets should have some limit on the earnings associated with very high offer caps to ensure scarcity pricing without price gouging. For instance, the Australian approach also includes a backstop feature called the cumulative price threshold (CPT), which caps the cumulative ﬁxed cost recovery over a 1 week period. When the limit is reached, the resource’s $AUS 10 000 offer cap drops to $AUS 50–$AUS100/MWh for no more than a week. The threshold that triggers the drop in the price cap has not been reached more than once in a year.48 In contrast, the ERCOT wholesale market has lower offer cap, but limits the amount a resource can capture on an annual basis to $175 000 per MW. When that limit is reached, a much lower offer cap applies for the remainder of the calendar year.49 Having such a cumulative cap in these markets with relatively high offer caps, however, could create a bright line, which enables pivotal suppliers to collect the allowed rents while staying within the allowed limits. Any market mitigation approach would need to address the problem of pivotal suppliers in the markets. Energy-only markets can work when two types of issues are addressed: transform the electricity market into a commodity market and address potential market power abuse through a combination of structural remedies (including transparency) and market monitoring tools. Concerns about free riders may require contracting or capital requirements (margin calls).

EBL

9.6. Transitional Mechanism Based on Energy-Only Markets with Contracting Obligations Resource adequacy mechanisms based on energy-only markets are premised on the notion that consumers are interested in buying deliverable energy, and hence generation capacity is valued based on its availability to produce energy at a given price. Under such 48

Authors’ communication with Peter Adams, Manager, Surveillance and Enforcement, Markets Branch, Australian Energy Regulator, February 1, 2005. 49 The lower offer cap in ERCOT market is $500/MWh or 50 times Houston Ship Channel Natural Gas Price Index, whichever is larger.

Resource Adequacy: Alternate Perspectives and Divergent Paths

355

mechanisms, generators can recover their investment costs either through very high spot prices that reﬂect scarcity rents or through long-term bilateral energy contracts with load-serving entities that specify a ﬁxed market-based energy price at levels signiﬁcantly below scarcity pricing but sufﬁciently high to cover generator’s amortized ﬁxed costs. These bilateral contracts are essentially energy market-based call options on generation capacity, with the capacity cost premium comparable to the premium on a ﬁnancial instrument that has a set strike price (i.e., the ﬁxed price for deliverable energy). Such mechanisms are contrasted in this chapter with resource adequacy approaches that are based on introducing an artiﬁcial capacity product for which there is no intrinsic demand. Capacity-based resource adequacy mechanisms can take the form of a direct capacity payment, as is the case in several Latin American countries, in European countries such as Italy and Spain, and in South Korea. Alternatively, the regulator can specify through regulatory ﬁat the needed quantity for capacity based on engineering considerations or specify an administrative demand curve for capacity (e.g., NYISO) from which the capacity price is inferred through a mitigated, auction-based procurement process. As discussed earlier, capacity payments, whether derived from an administrative demand curve or imposed directly, are aimed at restoring the missing money given the prevailing energy market and hence can complement any energy market design regardless of any pricing inefﬁciency. On the other hand, recovery of investment cost through energy-only adequacy mechanisms, whether the amount of installed capacity is implied by customer preferences for reliability or dictated by engineering standards and load forecasts, requires that energy prices be efﬁcient and be allowed to reﬂect scarcity and value of unserved load. In addition a functional market for risk must exist, either over the counter or centralized, which will enable generators and loads to trade investment and price risk. Such a market is essential for protecting customers and LSEs against extreme price excursions and for smoothing out boom–bust cycles that may adversely affect the cost of capital for investors. The restructuring of the electricity industry has focused primarily on improving efﬁciency of energy spot markets and mitigation of market power. Little attention has been devoted to facilitating markets for risk, which become crucial in an energy-only market where prices are allowed to reﬂect scarcity. The presumption is that fear of exposure to high spot prices will drive loads to seek cover of bilateral contracting arrangements. However, such a leap of faith leaves some open questions:

EBL

•

An LSE to contract forward for 100 percent of its forecasted peak load is not optimal given load uncertainty and potential load migration. • Credit constraints may prevent small retail energy providers from entering into contracts that extend far into the future. • There is always the option of bankruptcy which, like it or not, provides a hedge against extreme low-probability events. The question is then, whether investors in new generation are willing to take the risk of bankrolling an irreversible large-scale investment with a 30-year cost recovery horizon based on relatively short-term contracts and for only a portion of the capacity.50 • Can the system operator entrusted with maintaining reliability count on such investment or on load response in the face of a tight supply and demand balance? 50

This concern is not limited to the electric generation industry. For instance, the development of new sources of oil and gas involve large capital costs, with the output of these investments subject to volatile and uncertain prices over the life of the investment.

356

Competitive Electricity Markets

Currently, there are two markets in the US, MISO, and CAISO, where resource adequacy is assured through market-based bilateral contracting obligations imposed on loads or LSEs, described below. MISO: In February 2007, the Midwest ISO (MISO) ﬁled with FERC a resource adequacy proposal that relies on energy-only remuneration and contracting obligations imposed on LSEs. The proposal retains the $1000 per MWh energy offer caps during non-scarcity conditions (i.e., when reserves are not deployed for energy production). However, an administrative demand curve for reserves will be used to set reserve prices during scarcity conditions when operational reserves are deployed. These reserve prices will also be added to the energy clearing prices during such scarcity periods. The demand curve for reserves, which will be used in the day-ahead and real-time markets, will allow scarcity pricing to rise as high as $3500 per MWh, which is MISO’s estimate of VoLL. Real-time cooptimization of energy and reserves will be implemented to improve resource utilization during scarcity conditions. The state commissions within MISO would be expected to enforce a contracting requirement for all loads, both traditional cost-of service load and competitive retail loads, to ensure resource adequacy. The desirability of a “must-offer” availability requirement in day-ahead markets for contracted resources in MISO will be reviewed in the future. CAISO: In California, the unpleasant experience of the energy crisis led the California Public Utility Commission (CPUC) to take a proactive role in assuring that LSEs enter into bilateral contract that will ensure adequate resources to meet local reliability requirements determined by the California Independent System Operator (CAISO). The low offer cap of $400/MWh in the CAISO energy market and the absence of any capacity payments made bilateral contracting obligations an essential backstop mechanism, which would reduce the CAISO’s heavy reliance on RMR contracts to meet local reliability needs. In a series of orders, the CPUC set the ground rules for a resource adequacy requirement (RAR) program based on mandatory bilateral contracting obligations imposed on CPUCjurisdictional LSEs. The initial CPUC Resource Adequacy Orders were issued in 200451 where the CPUC established the LSE obligation framework, established qualifying capacity rules, and authorized a wide range of resource types. In a subsequent order in 200552 , the CPUC clariﬁed the notion of monthly capacity versus peak load, established required elements for standardized contracts, and clariﬁed the availability obligation to the CAISO of contracted generators. In 2006 the CPUC issued another order53 that resolved a number of regulatory uncertainties including treatment of forced outages versus scheduled outages, title clearing, creditworthiness, and the role of intermediaries. The order also modiﬁed the required elements of tradable, standardized capacity contracts, and authorized trading via bulletin boards or exchanges. The contracting obligations imposed on the LSEs are based on load forecasts developed by the California Energy Commission (CEC) and on local reliability needs determined by the CAISO. There are nine local reliability areas and each LSE is required to carry contracts or own generation covering 115–117% of its peak load share for the upcoming year in each local reliability area54 . The contract portfolios of the LSEs are subject to compliance veriﬁcation by the

EBL

51

CPUC Orders D.04-01-050 and D.04-10-035. CPUC Order D.05-10-035. 53 CPUC Order D.06-07-031. 54 To increase liquidity and facilitate the process the local reliability areas in northern California have been aggregated into two procurement areas. 52

Resource Adequacy: Alternate Perspectives and Divergent Paths

357

CPUC55 . In August 2006 the CPUC launched a proceeding aimed at determining whether there is a need for a centralized capacity market that will supplement or replace the current contract-based resource adequacy approach, which has several shortcomings.56 One of the problems with the current program is the overwhelming burden of compliance veriﬁcation due to the diversity of contracts. Furthermore, the dual role of the contracts addressing both price hedging and local reliability needs is problematic since from a ﬁnancial perspective it makes little sense to contract forward for over 100% of forecasted annual load. Optimal hedging of variable load would normally cover only a fraction of the peak load with forward ﬁxed price contracts while the remainder is served through spot market procurements. Covering 100% of the peak load with forward contracts is suboptimal, since under such a strategy the LSE is over hedged when demand is low. The LSE, as a result, would have to sell its excess energy in the balancing at a potential loss, since low demand in the face of oversupply is typically correlated with low spot prices. Call option obligations imposed on wholesale customers and LSEs are intended to overcome some of these difﬁculties and serve as a backup resource adequacy mechanism that corrects for possible failure in the market for risk with minimal intrusion in private risk management practices. They provide the same function as capacity products, but their pricing and performance obligation are linked to energy markets that must, therefore, be efﬁcient and unmitigated. Ideally, spot price exposure should motivate market participants to manage their exposure to price and volume risk through a portfolio of forward contracts, call and put options, and spot trading. However, market imperfections, spot price distortions due to regulatory interventions and reliability-motivated, out-of-market operator actions, result in over-reliance on the spot market by buyers, cost recovery shortfalls by producers, and under-investment. Such misallocation of risk in the electricity supply chain creates a vicious cycle where underhedging by load exposes customers to price risk, resulting in regulatory intervention to suppress price spikes. But such suppression of spot prices creates shortfalls in generators cost recovery and reduces incentives for investment, which in turn may result in higher spot price volatility due to shortages and threaten reliability. Standardized call option obligations can fulﬁll the role of a capacity mechanism in a transitional energy-only market, which has not matured to the point where voluntary contracting, ﬁnancial intermediation, and load response can meet local reliability needs.

EBL

9.6.1. Call options as mandatory load hedging Proposals for various forms of contracting obligations, in the form of physically covered call options, are described by Oren (2000, 2005a, b) and by Vázquez et al. (2002). Mandatory load hedging, although with no physical cover requirements, has also been proposed 55

In a 2004 long-term procurement decision (D. 04-04-003) the CPUC assigned the responsibility for new generation capacity buildup to the investor owned utilities (IOUs). The rule empowers the IOU to enter into contracts with new generation capacity and allocate the capacity cost component of the contracts, for up to 10 years, to all CPUC-jurisdictional LSEs in the IOUs service territory. The rule opens the door for an auction-based mechanism for reselling the energy tolling rights to a contract so that the capacity cost component can be inferred by deducting the energy value from the total contract cost. 56 The above summary is based on a presentation by Mike Jaske of the CEC at a CAISO Market Surveillance Committee Meeting held on August 8, 2006.

358

Competitive Electricity Markets

recently by Hogan (2005, 2006) in the context of an energy-only market proposal. A recent implementation of a call option approach with a central procurement auction in Brazil is described by Bezerra et al. (2006), and a similar approach has been proposed in Colombia by Cramton and Stoft (2007). As mentioned earlier, a mandatory contracting obligation has also been adopted in its recent MISO proposal to FERC and in California under a resource adequacy rule, which mandates that LSEs cover their peak load and reserve margin through bilateral contracts that are subject to veriﬁcation.57 Contracting obligations are intended to greatly reduce the potential for a boom–bust cycle by effectively imposing mandatory price insurance as a way to protect customers against high spot prices instead of artiﬁcially suppressing these prices. For such contract to be properly priced and to address the “missing money” problem, unhedged spot prices for energy and operating reserves must be allowed to fully reﬂect scarcity conditions. The purpose of mandatory call options is to restore the price protection offered to customers through price caps and ensure resource availability at these prices with minimal interference in the private risk management markets. This result can be achieved by setting the strike price for the mandatory call options sufﬁciently high (say $1000 per MWh, like the prevailing price cap in many markets). Under such a scheme any LSE has to cover its peak load and appropriate planning reserve requirements with the standardized backstop call options or any forward contract or call options of comparable duration and equal or lower strike price. The call options give the holder the right but not the obligation to obtain a ﬁxed amount of energy at the strike price, which can come from a generation resource or a curtailable load resource.58 Holding options allow LSEs to buy cheaper energy from the spot market when available, reducing the cost to consumer as compared to forward contract that entails a “must-take” obligation at the contract price. The LSE beneﬁts from holding call options, which offer customers the same protection as a price cap without locking them into a ﬁx price. Further price protection can be obtained through traditional private contracting. Call options can be obtained from generation resources with veriﬁable physical capacity to cover the contract or from interruptible ﬁrm load that by selling a call option commits to curtailment when the spot price reaches the strike price. The opportunity for load to sell call options effectively gives load an opt-out opportunity, since the revenue from selling the call option will offset their share of call option obligation cost that they may be subject to under an LSE. Resources that are not bound by call option contracts are allowed to offer their energy into the spot market at prices that exceed the strike price, while generators and interruptible loads that have offered call options are obliged to offer the corresponding energy into the spot market at the strike price or below and are liable, in case of non-performance, for the difference between the uncapped spot price and the strike price. Such liability automatically implements a non-performance penalty scheme based on market economics that keeps the customers unharmed.

EBL

57

CPUC orders D.04-01-050, D.04-10-035, D.05-10-035, D.06-07-031. An ISO might ﬁnd it prudent to require LSEs to procure some curtailable load resources with call options that have a strike price greater than $1000 per MWh in order to maintain system reliability during extreme weather events or random emergency conditions. Such an approach would allow the ISO to avoid mandatory load shedding by having loads that want to consume electricity during emergency conditions to pay other loads to curtail. The higher strike prices would reduce the upfront cost of procuring such voluntary load shedding. Such contracts could be procured annually rather than for 3 years to increase the potential range of load participation. 58

Resource Adequacy: Alternate Perspectives and Divergent Paths

359

9.6.2. Centralized versus decentralized market Backstop call option obligations can be met through bilateral contracting with generators or interruptible customers; however, in order to achieve the resource adequacy objective and provide incentives to new investment in generation capacity, it is necessary to deﬁne the call options with at least a 3 year forward lead time and an obligation duration of one to several years. The need for such a lead time has been recognized in other mechanisms mentioned earlier such as the PJM RPM, CRAM, and the ISO-NE FCM, as a way to allow new entrant with no “iron in the ground” to participate in the market and offer call options against capacity that they intend to build. Such a forward-looking approach is essential to mitigate market power of the incumbent generators in the call option market. The problem with forward-looking commitments is that load forecasts change, load migrates among LSEs, and small REPs may not have adequate credit to cover such long-term obligations. These difﬁculties may be addressed by creating a centralized market of last resort for call options. Such a market that will complement the traditional over-the-counter bilateral contract market can be managed by the ISO that will act as counterparty and assume all the credit risk. The ISO will centrally procure the backstop call options on behalf of the system load through an auction mechanism. Payment to the sellers of the options will be made at performance time and the cost will be recovered from the load through their LSE on a prorated basis. Hence load migration problem and load forecast errors are dealt with based on realized load. Self-provision in such a centralized market can be handled in two ways. One is to allow LSEs to submit proof of contracts they hold in lieu of their call option obligations. The better approach, however, is to have holders of qualifying contracts sell call options into the centralized auction covered by the contracts. The proceeds they will receive from such sold options will offset the charges for their obligation.

EBL

9.6.3. Deliverability issues To assure that the resource adequacy objective is achieved, i.e., sold call options translate into physical capacity, the call option must have physical cover, i.e., a seller of a call option must identify physical generation capacity or ﬁrm interruptible load or make a commitment to build the capacity within a certain time frame. The question that arises is how to guarantee that the capacity will be built where it is needed so that the energy it produces is deliverable. The approach that has been adopted in the Northeast ISOs is to deﬁne locational capacity markets that reﬂect transmission constraints. As a result, capacity prices in such systems will vary based on location. The call option approach enables such locational differentiation naturally in a locational marginal pricing (LMP)based system, since the ﬁnancial liability associated with a call option and the opportunity cost of committing to the strike price are determined by the difference between the LMP and the strike price so that the call option premium will vary by location. There is still the question, however, of how granular the call option obligation and the corresponding physical cover should be. The physical cover associated with the call options can have the same granularity as the ﬁnancial obligation or coarser. For example, the call option obligation can be zonal, implying that non-performance penalties will be based on the difference between the average zonal spot price and the strike price, but the physical cover can be anywhere in the system. In such a case, the generator selling the call option can cover its ﬁnancial risk with capacity anywhere in the system and shortterm transmission congestion rights between the location of the capacity and the location corresponding to the call option the generator sold.

360

Competitive Electricity Markets

9.7. Conclusions Workably competitive electricity markets with robust risk management schemes do not require regulatory-based capacity payments to generators. While this proposition was supported in the abstract by US academics and policymakers, the institutional realities of piecemeal electricity deregulation have made it seem unrealistic to many in the US. Working examples of such a market, such as those in Australia and Alberta, were either ignored or dismissed as anomalies. For the past 10 years, the worldwide debate about key elements of wholesale market design has been most intense in the US. The US debate has been dominated by two market design elements – real-time dispatch and resource adequacy. In both cases, PJM, the ISONE, and the NYISO took a strong, clear position in that debate – a combination of nodal real-time dispatch with ICAP markets (capacity resource adequacy mechanisms) – which carried over into the market design debates at FERC and in three other US ISOs: CAISO, ERCOT, and MISO. Stakeholders and the public utility commissions in CAISO, ERCOT, and MISO debated long and hard on the beneﬁts and drawbacks of the nodal/capacity market combination adopted in the eastern US. These markets adopted the nodal real-time dispatch but rejected the ICAP approach. The rejection of capacity mechanisms was the result of the failure of existing capacity mechanisms to meet their primary objective of encouraging new investment in generation and their troubled evolution into less market-friendly, more complicated forms. As a result, the momentum in resource adequacy has shifted toward the energy-only approach, something almost unimaginable a few years earlier. While the energy-only resource adequacy mechanism is the mechanism of choice from an economic theory perspective and has been proven feasible in various parts of the world, it may not be the right choice for everyone. Ultimately, the choice depends on existing circumstances of the market’s infrastructure, stakeholders’ preferences, and the extent to which the political environment can create the needed circumstances conducive to an energy-only approach. The success or failure of the energy-only resource adequacy mechanism at one or more ISOs will depend on the ability of those markets to put all the market design elements into place to transform their wholesale markets into electricity commodities markets, as has been done successfully in the Commonwealth markets such as Australia.

EBL

Acknowledgment Parviz Adib and Eric Schubert are former staff of the Public Utility Commission of Texas and Shmuel Oren has been Senior Advisor to the Commission since summer of 2000. The opinions expressed in this article are those of the authors and do not represent the opinion of the Public Utility Commission of Texas or its Staff. The authors would like to thank Jess Totten, Director of Competition Division at the Public Utility Commission of Texas, for his invaluable comments on earlier drafts of this chapter.

References Alberta Department of Energy (2005). Reﬁnement Options for Alberta’s Wholesale and Retail Electric Markets. Alberta, canada. 10 March. Barroso, L.A., Rudnick, H., and Hammons, T. (2006). Second wave of electricity market reforms in Latin America. Presented at IEEE PES Panel Session, IEEE General Meeting, Montreal, Canada.

Resource Adequacy: Alternate Perspectives and Divergent Paths

361

Bezerra, B., Augusto Barroso, L., Granville, S., et al. (2006). Energy call options auctions for generation adequacy in Brazil. Proceedings of the IEEE Annual Meeting, Montreal, Canada, June. Boiteaux, M (1960). Peak load pricing. J. of Bus., 33, 157–79. Charles Rivers and Associates [Ruff, Larry] (2004). A Transitional Non-LMP Market for California: Issues and Recommendations. Cramton, P. and Stoft, S. (2005). A capacity market that makes sense. Elec. J., 18(7), 43–54. Cramton, P. and Stoft, S. (2006). The convergence of market designs for adequate generating capacity. University of California Energy Institute Working Paper, 25 April. Cramton, P. and Stoft, S. (2007). Colombia ﬁrm energy market. Proceedings of the 40th Hawaii System Science Conference (HICSS40), Big Island, Hawaii, January. De Vries, L.J. (2004). Securing the public interest in electricity generation markets, the myths of the invisible hand and the copper plate. Ph.D. dissertation, Delft University of Technology, Faculty of Technology, Policy and Management, Available at: http://www.tbm.tudelft.nl/webstaf/laurensv/LJdeVries_dissertation.pdf. Electricity Consumers Resource Council [ELCON] (2006). Today’s Organized Markets – A Step Towards Competition or an Exercise in Re-Regulation? 4 December. Available at: http://www.elcon.org/ Documents/Publications/12-4piom.pdf Federal Energy Regulatory Commission (2002). Report of PJM Interconnection, LLC Re-Supporting Seasonal Capacity Commitment Structure. Docket No. EL01-63-001, 31 May. Federal Energy Regulatory Commission (2005). 2004 State of the Markets Report, June, Washington, D.C. Hogan, W. (2005). On an “energy only” electricity market design for resource adequacy. Working Paper, Center for Business and Government, 23 September. Hogan, W. (2006). On an “energy only” electricity market design for resource adequacy. Presentation at Eleventh Annual POWER Research Conference on Electricity Regulation and Restructuring, Berkeley, Ca, 24 March. Hurlbut D., Rogas, K., and Oren, S. (2004). Protecting the market from ‘Hockey Stick’ pricing: How the Public Utility Commission of Texas is dealing with potential price gauging. Elec. J., April, 17, 26–33. Joskow, P. (2005). Why capacity obligations and capacity markets. Available at: http://econwww.mit.edu/faculty/download_pdf.php?id=1175 Litvinov E, Yang J., and Zhen, T. (2004). Building locational-based ICAP market in New England. Presentation at the IEEE PES Summer Meeting, Denver, Colorado, 6–10 June. McNamara, R. (2006). Midwest ISO, resource adequacy in Midwest energy markets. Presentation at the Organization of MISO States, 8–9 May. New York Department of Public Service (2003a). Prepared Testimony of Raj Addepalli, Harvey Arnett and Mark Reeder. Regarding a Proposal by the NYISO Concerning Electricity Capacity Pricing, Albany, NY, 6 March. New York Department of Public Service (2003b). Proposal Regarding Resource Demand Curve. Albany, NY, 31 January. New York Independent System Operator Inc. (2004a). FERC Electric Service Tariff Article 5, Albany, NY. Available at http://www.nyiso.com/services/documents/ﬁlings/pdf/services_tariff/services_ tariff.pdf. New York Independent System Operator Inc. (2004b). ICAP Demand Curve Review. ICAP Working Group, PowerPoint Presentation by Levitan and Associates Inc., Albany, NY, 27 May. Oren, S. (2000). Capacity payments and supply adequacy in competitive electricity markets. In Proceedings of the VII Symposium of Specialists in Electric Operations and Expansion Planning (SEPOPE VII), Curitiba, Brazil, 21–6 May. Oren, S. (2005a). Ensuring generation adequacy in competitive electricity markets. In Electricity Deregulation: Choices and Challenges (M.J. Grifﬁn and S.L. Puller, eds). (BSSEPP) Bush School Series in the Economics of Public Policy, June. Oren, S. (2005b). Generation adequacy via call option obligations: safe passage to the promised land. Elec. J., 18, 28–42. PennFuture (2006). RPM ain’t R.I.P. PennFu. Newsl., 8(13), 25 October. Available at: http://www. pennfuture.org/media_e3_detail.aspx?MediaID=692&TypeID=3.

EBL

362

Competitive Electricity Markets

Schubert, E. (2005). An Energy-Only Resource Adequacy Mechanism. Public Utility Commission of Texas, Rulemaking Project No. 24255, 14 April. Schubert, E., Hurlbut, D., Oren, S., and Adib, P. (2006). The Texas energy-only resource adequacy mechanism. Elec. J., 19, 39–49. Shanker, R.J. (2003). Comments on standard market design, resource adequacy requirement. FERC Docket No. RM01-12-000, 10 January. Vázquez, C., Rivier, M., and Arriaga, I.P. (2002). A market approach to long-term security of supply. IEEE Trans. on Pow. Sys., 17, 349–57.

EBL

Chapter 10 The Evolution of PJM’s Capacity Market JOSEPH E. BOWRING PJM Interconnection, L.L.C.

Summary This chapter explains capacity market fundamentals and covers the evolution of the PJM capacity market rules from the introduction of formal capacity markets on 1 January 1999 through the current market structure, and concludes with a review of the Reliability Pricing Model (RPM) design which became effective on 1 June 2007.1 The PJM capacity market emerged in 1999 from the combination of a pre-market system that enforced capacity adequacy and state regulatory pressures designed to ensure access to required capacity for new retail market entrants in the context of retail deregulation. The evolution of PJM capacity market rules since 1999 and the redesign of capacity markets in 2007 reﬂected incremental market design improvements in response to price volatility and the exercise of market power. However, the capacity market design was not fundamentally modiﬁed to adjust to the integrations of new control areas in PJM between 2002 and 2005, in particular the locational differences in capacity market conditions that resulted. While the capacity market was not originally designed as a source of revenue, the redesign of the PJM capacity market in 2007 explicitly incorporated the central role of the capacity market as a source of revenue to generation owners and thus a source of incentives for both new investment and reinvestment in existing capacity. Net revenue from all PJM markets (energy, capacity, and ancillary services) over the eight years of PJM market history provided inadequate incentives to invest in generating capacity in PJM. The new RPM capacity market design which became effective on 1 June 2007 includes an annual market, a forward market, locational capacity markets, scarcity pricing of capacity via a deﬁned demand curve, clear links to the energy and ancillary services markets, incentives to provide energy reliability, and clear market power rules including a must offer requirement. The RPM design is a clear improvement over the prior design, but further improvements are required. The impact of the RPM design on investor expectations and investment in existing and new capacity remains to be seen.2

EBL

1

117 FERC ¶ 61,331 (2006). The material in this chapter draws on the work of the PJM Market Monitoring Unit (MMU) and particularly the MMU’s 2006 State of the Market Report. I want to acknowledge the fact that the 2006 State of the Market Report is based on substantial contributions by all members of the MMU, 2

363

364

Competitive Electricity Markets

10.1. Introduction An essential feature of wholesale power markets is that they are required to meet externally imposed reliability standards. These reliability standards are enforced through a requirement to maintain a target level of installed or unforced capacity.3 This exogenous reliability requirement must be met, whether capacity markets are incorporated or whether the market design is an energy-only design. While it is clear how to incorporate the target level of capacity in a capacity market, the same is not true for an energy-only market. It is as difﬁcult or more difﬁcult to manage energy prices so as to produce the target level of capacity in a market where demand remains extremely price inelastic and the physical ability does not yet exist for individual customers to signal or to receive their desired level of reliability.4 The right level of reliability will not emerge automatically from an energy-only market design any more than it will from a design that includes a capacity market.5 The actual level of capacity results from investor decisions which are a function of investor expectations about the level and volatility of net revenues that will result from the market design. In an energy-only market designed to meet a target level of reliability, net revenues result from the management of scarcity pricing. Administrative scarcity pricing levels must be set so as to produce the required level of net revenues from the energy market while adjusting for variations in demand and supply in real time, addressing locational issues and preventing the exercise of market power. Given that maintaining a target level of capacity will tend to reduce energy price levels and volatility, the required scarcity prices in an energy-only market are likely to be extremely high, to be relatively infrequent, and to occur for unpredictable periods. To the extent that scarcity pricing levels require ongoing intervention and modiﬁcation to achieve target reliability levels, regulatory risk is also introduced. Investor expectations are likely to incorporate the resultant volatility and regulatory risk in the form of a risk premium. Even with a capacity market, energy market design must permit scarcity pricing when such pricing is consistent with market conditions and constrained by reasonable rules to ensure that market power is not exercised. Scarcity pricing is part of an appropriate incentive structure facing both load and generation owners in a working wholesale electric power market design. Scarcity pricing is a key link between energy and capacity markets. With a capacity market design that incorporates scarcity rents in the energy market, scarcity pricing can be a mechanism to increase reliance on the energy market as a source of revenues and incentives in a competitive market, without sole reliance on the energy market to signal the need for the target level of capacity.

EBL

including Kevin Bazar, Jerry Bell, Tom Blair, Susan Cawley, Bridgid Cummings, Andrew Engle, Beatrice Gockley, Howard Haas, Ellen Krawiec, Mark Million, Grace Niu, Jack O’Neill, Frank Racioppi, Paul Scheidecker and Tom Zadlo. 3 Unforced capacity is installed capacity adjusted for the relevant forced outage rate. 4 Advanced metering would allow customers to see the real-time price, to react to that price and to receive the beneﬁts of that reaction, thus providing demand elasticity. Advanced metering could also allow service to be interrupted to individual customers, allowing differentiated charges for varying levels of reliability. 5 This is a critical point which is made in two recent papers: Cramton, P. and Stoft, S. (2006). The convergence of market designs for adequate generating capacity with special attention to the CAISO’s resource adequacy problem. April; and Joskow, P. and Tirole, J. (2007). Reliability and competitive electricity markets. RAND J. Econ. Spring, Vol. 38–1, pp. 60–84.

The Evolution of PJM’s Capacity Market

365

10.1.1. PJM overview PJM Interconnection, L.L.C., operates a centrally dispatched, competitive wholesale electric power market that in 2006 had an average installed generating capacity of 162 571 megawatts (MW), a peak load of 144 644 MW and more than 450 market buyers, sellers, and traders of electricity in a region, including more than 51 million people in all or parts of Delaware, Illinois, Indiana, Kentucky, Maryland, Michigan, New Jersey, North Carolina, Ohio, Pennsylvania, Tennessee, Virginia, West Virginia, and the District of Columbia (see Fig. 10.1). As part of that function, PJM coordinates and directs the operation of the transmission grid and plans transmission expansion improvements to maintain grid reliability in this region. PJM operates a day-ahead energy market, a real-time energy market, a capacity market, a regulation market, a synchronized reserve market, and a ﬁnancial transmission rights (FTR) market. PJM introduced energy pricing with market-clearing nodal prices and cost-based offers on 1 April 1998, and market-clearing nodal prices with market-based offers on 1 April 1999. PJM introduced a daily capacity market on 1 January 1999, and monthly and multimonthly capacity markets in mid-1999. PJM implemented an auction-based FTR market on 1 May 1999. PJM implemented a day-ahead energy market and a regulation market on

EBL

Fig. 10.1. PJM’s footprint and its zones. PJM Market Monitoring Unit (2006). 2006 State of the Market Report. PJM Geography, Vol. II, Appendix A.

366

Competitive Electricity Markets

1 June 2000. PJM added a market in spinning reserve on 1 December 2002. PJM introduced an auction revenue rights (ARR) allocation process effective 1 June 2003.6 10.1.2. Brief history of PJM integrations and capacity markets Initially, PJM was comprised of the Mid-Atlantic Region, including its 11 zones.7 From 1 January 1999 through 31 March 2002, PJM operated a single capacity market. From 2002 through 2005, PJM integrated six additional control zones.8 Effective 1 April 2002, PJM integrated the Allegheny Power Company (AP) Control Zone.9 The capacity market rules for the AP Control Zone, also termed the PJM Western Region, were different but closely related to the PJM capacity market rules. On 1 June 2003, the PJM Western Region capacity market was incorporated into the PJM capacity market. Effective 1 May 2004, PJM integrated the Commonwealth Edison Company Control Area (ComEd). From 1 June 2004 through 31 May 2005 PJM operated a separate ComEd capacity credit market. Effective 1 October 2004, PJM integrated the American Electric Power Control Zone (AEP) and The Dayton Power & Light Company Control Zone (DAY). Effective 1 January 2005, PJM added the Duquesne Light Company (DLCO) Control Zone. Effective 1 May 2005, PJM integrated the Dominion Control Zone. After 1 June 2005, PJM operated a single capacity market throughout its footprint. 10.2. Emergence of a Capacity Market in PJM A capacity market was not included in the ﬁling by the PJM member companies that created the PJM markets.10 It was apparently anticipated that energy markets would provide adequate revenue to generation owners. Nonetheless, there was a system of capacity obligations that predated organized wholesale markets in PJM, which continued with the introduction of wholesale markets, and that presaged a capacity market. Well prior to organized markets, PJM and its members relied upon capacity obligations to ensure reliability and to allocate the costs of that reliability. The eight original PJM members determined their loads and related capacity obligations on an annual basis. Capacity obligations were based on forecast load plus an agreed-upon reserve margin that was designed to achieve the desired reliability objective. The PJM agreements used capacity obligations to deﬁne and enforce the requirement that member

EBL

6

See also PJM Market Monitoring Unit. (2006). 2006 State of the Market Report. PJM Market Milestones, Vol. II, Appendix B. 7 The Mid-Atlantic Region is comprised of the Atlantic Electric Company Control Zone (AECO), the Baltimore Gas & Electric Control Zone (BGE), the Delmarva Power & Light Control Zone (DPL), the Jersey Central Power & Light Company Control Zone (JCPL), the Metropolitan Edison Company Control Zone (Met-Ed), the PECO Energy Company Control Zone (PECO), the Pennsylvania Electric Company Control Zone (PENELEC), the Pepco Control Zone (PEPCO), the PPL Electric Utilities Corporation Control Zone (PPL), the Public Service Electric and Gas Company Control Zone (PSEG), and the Rockland Electric Company Control Zone (RECO). 8 See PJM Market Monitoring Unit. (2006). 2006 State of the Market Report. PJM Geography, Vol. II, Appendix A. 9 Zones, control zones and control areas are geographic areas that customarily bear the name of a large utility service provider operating within their boundaries. Names apply to the geographic area, not to any single company. 10 PJM Supporting Companies (1997). Request for authorization to engage in market-based transactions, Docket No. ER97-3729-000, 14 July.

The Evolution of PJM’s Capacity Market

367

utilities each contribute an appropriate share of capacity to the pool or pay a capacity deﬁciency charge if they did not. As a result of the lumpy nature of investments in generation, at times some companies could be long capacity and some could be short capacity. The PJM agreements included a mechanism for the short capacity companies to compensate the long capacity companies. Compensation was in the form of a capacity-deﬁciency charge based on the levelized cost of a peaking unit, minus expected net energy market revenues. A bilateral market in capacity obligations developed in which PJM members exchanged capacity obligations at prices that, in the early 1990s, were generally less than the deﬁned deﬁciency price, reﬂecting the excess of capacity supply over demand at the time. Prior to the creation of wholesale power markets, PJM members were all vertically integrated utilities subject to regulation by state public utility commissions. Such rate base, rate-of-return regulation provided direct incentives to build new capacity in order to meet the regulatory obligation to serve load. Generation-owning utilities were effectively guaranteed that their revenues would be adequate to cover the costs of new and existing capacity. The cost of capacity obligations was borne by members and recovered from loads under approved retail rates. The system of PJM capacity obligations established administratively deﬁned capacity targets designed to achieve a target level of reliability. When combined with state regulatory incentives to maintain adequate capacity, the result was a reliable pool with capacity and energy adequate to serve load and revenues adequate to provide compensation to generation owners in the period prior to the introduction of markets.

EBL

10.2.1. Implementation of a capacity market

The capacity market was implemented by PJM in response to a need created by the early stages of retail restructuring rather than in response to developments in the wholesale market. In the PJM area, the ﬁrst steps toward retail restructuring took place in Pennsylvania and were followed by restructuring in other states within PJM.11 Retail restructuring created the opportunity for new entrants to compete with incumbent utilities to serve retail loads. The new entrants, competing with vertically integrated utilities that owned generation assets in PJM, needed a competitive way to acquire capacity to meet reliability obligations, based on PJM capacity rules, associated with newly acquired retail loads. The new entrants needed a ﬂexible way to acquire capacity in relatively small increments and needed the ability to buy and sell capacity as loads switched suppliers. Relying solely on voluntary bilateral sales and purchases was likely to facilitate the exercise of market power and serve as a barrier to entry, particularly when the new entrant had to purchase capacity from the utility with which it was competing for retail load. As retail competition extended more widely over the PJM footprint, existing utilities also needed a way to sell capacity in excess of current requirements if load were lost to new competitors. Based, in part, on complaints by new entrants that the owners of capacity were unwilling to sell capacity to them at competitive prices, the Pennsylvania Public Utility Commission (PAPUC) issued an Interim Order on 17 September 1998, directing electric utilities under its jurisdiction and involved in PJM markets to immediately release for sale installed capacity at a price of $19.72 per KW-year ($54.03 per MW-day), which was the capacity 11

“Electricity Generation Customer Choice and Competition Act,” 66PA.C.S Chapter 28, as amended 20 November 1996. Available at http://www.puc.state.pa.us/electric/pdf/HB1509P4282.pdf

368

Competitive Electricity Markets

value used in the PAPUC’s stranded cost orders. The PAPUC stated that the Interim Order “is necessitated by recent actions of the holders of capacity in PJM, in which a few entities hold virtually all of the installed capacity necessary for every market participant to serve end use customers in the PJM area. The inexplicable market failure in installed capacity in PJM appears to present a situation in which capacity holders are acting in a monopolistic or oligopolistic fashion in an attempted exercise of market power and to resist market entry by new competitors.”12 In response, PJM ﬁled a proposal with FERC to create a capacity market on 14 October 1998 and ﬁled proposed modiﬁcations with FERC on 19 November 1998 to, among other things, make capacity market participation mandatory for an introductory period of ﬁve months.13 Based, in large part, on the creation of PJM capacity markets, the PAPUC reached settlement agreements with three electric utilities that ﬁled appeals of the Interim Order.14 Effective 1 January 1999, PJM implemented a mandatory capacity market. To meet the need for ﬂexibility created by the ability of load to switch suppliers, the capacity market was structured as a daily market. For an initial ﬁve-month period suppliers were required to offer available capacity daily and load serving entities were required to buy capacity to meet daily reliability requirements. There was no requirement to own capacity for longer than a day. At the request of regulators and others, the mandatory offer requirement was subsequently extended for an additional year. The capacity market did not create the demand for capacity, which was created by the same PJM rules that predated the creation of the market. The capacity market was established as a mechanism to trade capacity so that rights to existing capacity could follow load, to facilitate retail competition.15

EBL

10.3. Capacity Market Fundamentals In PJM, the demand for capacity is administratively determined based upon the application of an agreed upon set of reliability criteria. While the rules are complex, the key point is that the capacity requirement is designed to achieve a target level of reliability. The total demand for capacity is a deﬁned annual MW level and virtually completely priceinelastic. The PJM Reliability Assurance Agreement (RAA) includes rules that determine the total capacity obligation for PJM, the resultant total demand for capacity, and rules for allocating the total capacity obligation to individual load serving entities (LSEs).16 The total demand for capacity, net of Active Load Management (ALM), can be met through self-supply, bilateral contracts, or through the capacity market. 12

Pennsylvania Public Utility Commission (1998). Interim Order. Docket No. I-00980078 et al., 17 September. 13 PJM Interconnection, L.L.C. (1998). Filing of amendments to Amended and Restated Operating Agreement. Docket ER99-196-000, 14 October; PJM Interconnection, L.L.C. (1998). Amendment to Filing Regarding Capacity Credit Markets. Docket ER99-196-000, 19 November. 14 See also “Pennsylvania Public Utility Commission 1998/1999 Annual Report” and “Annual Report of the Pennsylvania Ofﬁce of Consumer Advocate Fiscal Year 1998–1999.” 15 The PJM capacity market is a capacity credit market which permits generation owners to retain full ownership and control of generating assets while selling the rights to capacity, capacity credits, for purposes of meeting the reliability requirements of load serving entities. 16 There are multiple RAAs, corresponding to broad geographical areas that were added to PJM through the integration process.

The Evolution of PJM’s Capacity Market

369

In the PJM capacity market, forecast peak loads are adjusted to establish capacity obligations.17 The adjusted forecast peak-load value is multiplied by the forecast pool requirement (FPR) to determine the unforced capacity obligation for PJM. The FPR is equal to one plus a reserve margin, multiplied by the PJM unforced outage factor. An LSE’s unforced capacity obligation for a zone is based on its customers’ aggregate share of the prior summer’s weather-normalized zonal peak load multiplied by zonal scaling factors and the FPR.18 The LSE’s zonal obligation may be further adjusted for ALM credits. The FPR is set for each planning period which commences every 1 June. Each LSE must own or purchase capacity greater than or equal to its capacity obligation. Under the pre-RPM rules, the net capacity position of every LSE is calculated daily when its capacity resources are compared to its capacity obligation. If an LSE’s capacity resources are less than its obligation, the LSE is deﬁcient. Deﬁcient entities must contract for capacity resources to satisfy their deﬁciency. Any LSE that remains deﬁcient must pay a penalty equal to the capacity deﬁciency rate (CDR), which was in the range of $165 to $180 per MW-day during the 1999 through 2007 period. After the 2001 effective date of the interval market, the penalty is the CDR multiplied by the number of days in an interval, where intervals are from three to ﬁve months in length.

10.3.1. Capacity resources (supply) Capacity resources are deﬁned as MW of net generating capacity meeting PJM-speciﬁc criteria. Capacity resources may be located within or outside of PJM, but they must be committed to serving load within PJM. All capacity resources must pass tests regarding the capability of generation to serve load and the deliverability of the energy to PJM load, which requires adequate transmission service.19 The supply of capacity is a function of physical capacity in PJM and capacity imports and exports. Capacity imports and exports are, in turn, a function of energy and capacity prices in external markets; energy and capacity prices in PJM markets; and transmission service availability and price. Capacity imports must be from speciﬁc generating units and sellers must have ﬁrm transmission from the identiﬁed units to PJM. Generators evaluate the opportunities to sell capacity on a continuing basis and sell capacity into the most proﬁtable market or set of markets. The sale of a generating unit as a capacity resource within PJM entails speciﬁc requirements. The sale and purchase of capacity resources without speciﬁc requirements would be just a cost allocation mechanism rather than a market mechanism designed to achieve reliability in a least cost manner. The purpose of the capacity requirement is to create a reliable, competitive energy market. Energy is the product customers require and the goal is reliable delivery of energy. Thus, capacity market rules must establish a strong link to performance in the energy market. Under the rules in place prior to RPM, the requirements associated with being a capacity resource are: day-ahead energy market offer requirement; energy recall right; deliverability; and generator outage reporting. The RPM rules provide additional performance incentives.

EBL

17

See “PJM Manual 17: Capacity Obligations,” Revision 6 (1 June 2005) (105 KB). 18 Zonal scaling factors are applied to historical peak loads to produce forecasted zonal peak loads. 19 See PJM (2004). Reliability Assurance Agreement, Capacity Resources, 17 May, p. 2, at: (344 KB).

370

Competitive Electricity Markets

PJM rules specify that when a generation owner sells a unit as a capacity resource, the seller is contractually obligated to offer their output into PJM’s day-ahead energy market. While the tariff rules do not specify the price, the only interpretation of this requirement with any meaning is that the offers must be made at a competitive price. When LSEs purchase capacity, they ensure that resources are available to provide energy on a daily basis, not just in emergencies. Since day-ahead offers are ﬁnancially binding, PJM capacity resource owners must provide the offered energy at the day-ahead clearing price, if the offer is accepted in the day-ahead energy market. The must offer requirement establishes a link between the capacity resource and actual offers to provide energy to the energy market. Thus, buyers and sellers in the PJM energy markets know that the markets will clear based on the inclusion of competitive energy offers from all capacity resources and those offers support the competitiveness and reliability of the energy markets. PJM rules specify that when a generation owner sells a unit as a capacity resource, the seller is contractually obligated to allow PJM to recall the energy generated by that unit if the energy is sold outside of PJM. This right enables PJM to recall energy exports from capacity resources when it invokes emergency procedures.20 The recall right establishes a link between the capacity resource and actual delivery of energy when it is needed in an emergency. An emergency can be thought of as approximating a scarcity condition, although there is substantial discretion involved in the recall decision. Thus, PJM can call upon energy from all capacity resources to serve load within the Control Area. When PJM invokes the recall right, the energy supplier is paid the PJM real-time energy market price. To qualify as a PJM capacity resource, energy from the generating unit must be deliverable to load in the PJM Control Area. Deliverability means that adequate transmission capability must exist or be constructed to ensure that energy from the capacity resource can be delivered to the PJM market. Capacity resources must be deliverable to the total system load, including portion(s) of the system that may have a capacity deﬁciency.21 In addition, for external capacity resources, capacity and energy must be deliverable to PJM through ﬁrm transmission service. Owners of PJM capacity resources are required to submit outage data to PJM.22 This data is important both for ongoing reliability assessments and for determining the actual, available level of unforced or net capacity. Each of these requirements is signiﬁcant, both for the energy markets and for reliability, but they are not all enforced in a meaningful way in PJM. While all capacity resources must be offered into the day-ahead energy market, there are a variety of ways to avoid compliance and there are no tariff rules to address these issues. There are no explicit tariff rules preventing withholding in the day-ahead market. Economic withholding occurs when a generation owner submits an offer above the competitive level for a unit. Withholding also occurs when a generation owner designates part or all of a unit as maximum emergency status, meaning that part or all of the unit would be available in an emergency but does not have an economic offer. It would be preferable to have explicit tariff language requiring competitive offers in the day-ahead market in order to make the must offer requirement meaningful. While

EBL

20

See “PJM Manual 13: Emergency Operations,” Revision 19 (1 October 2004). Available at: (461 KB). 21 The deﬁnition of deliverable is in “Reliability Assurance Agreement,” Schedule 10 (17 May 2004), p. 52 (344 KB). 22 See PJM (2004). Reliability Assurance Agreement, Schedule 12, 17 May, p. 57, at: (344 KB).

The Evolution of PJM’s Capacity Market

371

the energy from capacity resources is subject to recall during an emergency, PJM only rarely exercises this recall right. There are no tariff rules governing the exact terms under which energy will be recalled and the actual recall and timing and nature of recalls are left to the discretion of PJM operators in real time. It would be preferable to have clear rules governing recalls. While the deliverability rules have resulted in the construction of signiﬁcant transmission upgrades, they have not meant that all capacity is literally deliverable throughout the system. While owners of PJM capacity resources are required to submit outage data to PJM, the rules governing the classiﬁcation of outages remain unclear and subject to gaming.

10.3.2. Capacity market structure and market power The capacity market, by design, is always tight, in the sense that total supply is generally only slightly larger than demand. This is true for the historical capacity market design as well as the RPM design. The market may be long at times, but that is not the equilibrium state. Capacity in excess of demand is not sold and, if it does not earn or does not expect to earn adequate revenues in other markets, will retire. The demand for capacity includes expected peak load plus a reserve margin. Thus, the reliability goal is to have total supply equal to or slightly above the demand for capacity. Demand is almost entirely inelastic because the market rules require loads to purchase their share of the system capacity requirement. The result is that any supplier that owns more capacity than the difference between total supply and the deﬁned demand is pivotal. In PJM, in 2006, the excess supply was 9531 MW. There were four individual suppliers who owned more than 9531 MW of capacity and who were, therefore, each pivotal on a stand-alone basis. In other words, the market design for capacity leads, almost unavoidably, to structural market power in the capacity market. This is not surprising, given that the capacity market is the result of a regulatory/administrative decision to require a speciﬁed level of reliability and the related requirement that all LSEs purchase a share of the capacity required to provide that reliability. But, it is important to keep these basic facts in mind when designing and evaluating capacity markets. The capacity market is unlikely ever to approach a competitive market structure in the absence of a substantial and unlikely structural change that results in much greater diversity of ownership.23 In the capacity market, as in other markets, market power is the ability of a market participant to increase the market price above the competitive level or to decrease the market price below the competitive level. In order to evaluate whether actual prices reﬂect the exercise of market power, it is necessary to evaluate the competitive market price. A competitive offer in a capacity market is a function of the marginal cost of capacity, assuming no scarcity and accounting for opportunity costs where appropriate, less net revenues from the energy and ancillary services markets. The marginal cost of capacity is, in turn, a function of the time period for which the capacity is offered as well as of the alternative opportunities available to the generation owner. If an owner is considering whether to sell an existing capacity resource for a year, marginal cost includes the incremental cost of maintaining the unit for that year (going forward cost or avoidable cost) so that it can qualify as a capacity resource and any relevant opportunity cost, less

EBL

23

The market could be competitive if there were many more suppliers and all were relatively small compared to the size of the market and the level of excess capacity, but this is unlikely to occur.

372

Competitive Electricity Markets

unit-speciﬁc net revenues from the energy and ancillary services markets.24 If an owner is considering whether to construct a new resource and offer it to the market, marginal cost would include the full levelized annual cost, less expected net revenues from the energy and ancillary services markets. If an owner is considering whether to sell an existing capacity resource for a day, the only relevant cost is the opportunity cost for the day. The opportunity cost associated with the sale of a capacity resource in PJM is a function of the expected difference between energy and capacity prices in external markets and energy and capacity prices in PJM, and the expected probability that the energy will be recalled. Given the basic features of market structure in the PJM capacity market, including the existence of pivotal suppliers, inelastic demand, tight supply–demand conditions, the relatively small number of non-afﬁliated LSEs, the capacity-deﬁciency penalty structure facing LSEs, supplier knowledge of the penalty structure, and supplier knowledge of aggregate market demand, if not individual LSE demand, the potential for the exercise of market power is high. Market power is endemic to the PJM capacity market. There were no explicit rules providing for the identiﬁcation or mitigation of market power in the capacity market from 1999 until the introduction of the RPM rules effective 1 June 2007. 10.4. Market Evolution 10.4.1. Summer 2000 In 1999, capacity market prices averaged $52.86/MW-day over all capacity markets including daily, monthly, and multi-monthly markets. Monthly capacity market prices averaged $70.66/MW-day. Daily capacity market prices averaged $3.63/MW-day while the highest daily market price was $55/MW-day.25 In 2000, both daily and monthly capacity prices remained at or below $40/MW-day for the ﬁrst ﬁve months. However, on 1 June 2000 the daily price rose to $350.43/MWday, the highest level since the market was introduced. On 2 June, the daily price fell to $174/MW-day and on 3 June, the daily price was about $177, high by historical standards, where it remained for the balance of June. The prices and behavior in the capacity markets for June were consistent with the underlying supply and demand fundamentals, and there was no evidence of an exercise of market power. The results were directly related to speciﬁc features of the capacity market rules, in particular the deﬁnition of capacity as a daily product. Daily capacity markets created incentives to export capacity when the daily forward price differential between external and internal markets increased. Effective 1 June, a signiﬁcant amount of demand shifted into the daily market, probably as a result of the persistently low daily prices that had been experienced. At the same time, supply decreased as a result of increased exports in response to price differentials and the associated increase in the opportunity costs of selling capacity in the PJM daily markets. On 1 June 2000, for the ﬁrst time since the introduction of the PJM daily capacity markets on 1 January 1999, the total demand for daily capacity in the PJM markets exceeded the

EBL

24 This marginal cost, termed avoidable cost in the RPM tariff, does not include ﬁxed costs such as depreciation and return and, unless appropriate for a speciﬁc unit, does not consider retirement as the benchmark. The avoidable cost is the cost that could be avoided by not operating, or mothballing, the unit for a single year. 25 All ﬁgures are in unforced capacity terms, unless otherwise noted.

The Evolution of PJM’s Capacity Market

373

total supply of daily capacity. As a result, the PJM system as a whole was short of capacity when compared to peak season reliability requirements. The exports persisted throughout the summer, leaving PJM short of capacity on 16 peak season days. As a result of actual demand and supply conditions in the energy market, the system remained reliable. 10.4.2. First quarter 2001 A seller in the PJM daily capacity market took actions that resulted in an increase in the market price above the competitive level for a portion of the period from 1 January to 30 April 2001. The PJM MMU concluded that a single entity, acting unilaterally, exercised market power.26 The seller was individually pivotal and had the ability to determine the market price. The seller’s available capacity exceeded the difference between the daily demand for capacity and the available capacity of all other sellers. LSEs had to buy capacity from this seller to cover their obligations. If these LSEs became deﬁcient, they would be required to pay the capacity deﬁciency rate (CDR) of $177.30/MW-day. The allocation rules provided an incentive to holders of unsold capacity to offer it for sale at a price greater than or equal to the capacity deﬁciency rate. If LSEs did not purchase capacity at a $177.30 per MW-day, they would be deﬁcient and pay $177.30 per MW-day as the deﬁciency rate. The deﬁciency revenues were distributed to the withholder of the unsold capacity resources under the allocation rules in effect at the time. 10.4.3. Modiﬁed market design

EBL

In response to the observed market results, in March 2001, PJM ﬁled a proposal for change in the method used to allocate capacity deﬁciency revenues with FERC.27 The rule change expanded the pool of recipients of deﬁciency charge revenues to include all LSEs that met their obligation. In circumstances where an owner is pivotal, the amendment partially mitigated the owner’s ability to set the market price at the CDR. PJM also ﬁled in 2001 with FERC a proposal to implement an interval market which gave LSEs an incentive to meet their obligation to serve load on an interval basis and gave capacity owners a corresponding incentive to sell capacity on an interval basis. These changes became effective 1 July 2001.28 The new rules resulted in an improved alignment between market incentives and system reliability requirements, as well as a reduction of incentives to exercise market power in the daily capacity credit markets. The new rules also increased the time period for which an LSE must meet its capacity obligations from daily to a seasonal interval (ranging from three to ﬁve months), increased the deﬁciency charge provisions to provide for an interval penalty rather than a daily penalty, and required generation owners to commit excess capacity to PJM for an entire interval in order to participate in the allocation of any capacity deﬁciency charge revenues. 26

See PJM Market Monitoring Unit’s 2001 State of the Market Report (June 2002) and the PJM Market Monitoring Unit’s “Report to Pennsylvania Public Utility Commission, Capacity Market Questions” (November 2001) (http://www.pjm.com/markets/market-monitor/downloads/mmureports/20011121-rpt-pa-utility.pdf). 27 The FERC accepted the rule change effective 1 June 2001. PJM Interconnection, L.L.C., 95 FERC ¶ 61,175 (2001). 28 PJM Interconnection, L.L.C., 95 FERC ¶ 61,330 (2001).

Competitive Electricity Markets

374

As a result of the rule changes and changes in underlying market conditions, LSE reliance on the daily market decreased and prices declined in the daily, monthly, and multi-monthly markets to levels more consistent with a competitive outcome. 10.4.4. 1999–2006 Summary

EBL

9,000

Daily CCM

Monthly/multimonthly CCM Daily weighted-average price

$250

Monthly/multimonthly weighted-average price

8,000

$200

7,000 6,000

$150

5,000 4,000

$100

3,000 2,000

$50

1,000 0

Weighted-average capacity clearing price ($/MW-day): Lines

10,000

$0 Jun-99 Sep-99 Dec-99 Mar-00 Jun-00 Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02 Jun-02 Sep-02 Dec-02 Mar-03 Jun-03 Sep-03 Dec-03 Mar-04 Jun-04 Sep-04 Dec-04 Mar-05 Jun-05 Sep-05 Dec-05 Mar-06 Jun-06 Sep-06 Dec-06

Average daily capacity credits (Unforced MW): Bars

PJM markets, and therefore PJM capacity markets, grew substantially between 1999 and 2006. Average installed capacity in PJM in 1999 was 57 071 MW and in 2006 was 162 571 MW.29 The increase was almost exclusively the result of the integration of additional control areas between 2002 and 2005. While capacity supply exceeded demand in every year, that excess decreased to an average of only 695 MW in the summer of 2004 before increasing to about 10 000 MW with the integration of AEP and DPL. Prices in the capacity markets were higher in the period from 1999 through 2002, lower and cyclical through September 2004, and then quite low and stable (less than $10 per MW-day) in 2005 and 2006, reﬂecting that increase in excess supply. (See Fig. 10.2.) Market results were generally competitive with the exception of the ﬁrst half of 2001. Purchases in the PJM capacity markets accounted for a small but fairly stable share of total capacity obligations, varying from 2.6% in 1999 to 6.6% in 2006. Retail competition, as measured by shares of load obligation, reﬂected relatively low levels of market participation by companies other than PJM utilities and their afﬁliates. PJM utilities or electricity distribution companies (EDCs), together with their generating and marketing afﬁliates, served 87.6% of load obligation in 2006 with the balance served by EDCs not based in PJM, 5.6%, and non-EDCs not based in PJM, 6.8%.30

Fig. 10.2. PJM Daily and Monthly/Multimonthly CCM: June 1999 to December 2006. See PJM Market Monitoring Unit’s 2006 State of the Market Report, p. 227. CCM refers to the PJM capacity credit market because capacity credits are traded rather than physical unit capacity.

29

The 1999 data is for the 1 June through 31 December period. See PJM Market Monitoring Unit’s 1999 State of the Market Report and 2006 State of the Market Report. 30 PJM Market Monitoring Unit’s 2006 State of the Market Report, pp. 213–216.

The Evolution of PJM’s Capacity Market

375

10.5. The Performance of Markets: Net Revenue The ultimate test of a competitive wholesale power market design is whether it provides incentives to invest in capacity at a level consistent with the desired level of reliability and market participants invest based on incentives endogenous to the competitive market design rather than on reliance on the potential or actual exercise of market power. The expectations of investors as to the probability of achieving a target rate of return drive investment and are, in turn, formed by the design and results of the markets. In this context, net revenue is the single best measure of the results of wholesale power markets and of the potential role of capacity markets. Net revenue is an indicator of generation investment proﬁtability, and thus is a measure of overall market performance as well as a measure of the incentive to invest in new generation to serve PJM markets. Net revenue is the amount that remains, after variable costs have been subtracted from gross market revenue, to cover ﬁxed costs including a return on investment, depreciation, taxes, and ﬁxed operation and maintenance expenses. Net revenue quantiﬁes the contribution to ﬁxed costs received by generators from PJM energy, capacity, and ancillary service markets and from the provision of black start and reactive services. The market design goal is that on average, net revenue from all sources will cover the ﬁxed costs of investing in new generating resources when needed to meet reliability objectives, including a competitive return on investment. Wholesale energy markets, like other markets, are cyclical, and actual results can be expected to vary from year to year. When the markets are long, prices will be lower and when the markets are short, prices will be higher. Nonetheless, if potential investors do not expect that net revenues will provide a competitive return, they will not invest. While the reasons are complex, the facts are clear. In PJM, net revenue from the energy, capacity, and ancillary services markets was, in general, well below the replacement cost of capacity for combustion turbines (CT), combined cycle units (CC), and coal plants (CP) for the eight-year period from 1999 through 2006, regardless of analytical approach.31 (See Table 10.1.) Net revenues from the capacity market declined in part as a result of the modiﬁed supply demand fundamentals produced by the integrations in PJM. The increase in excess supply also put downward pressure on net revenues from the energy market. As a result of the integrations, the net revenue experience of PJM during this period does not reﬂect a closed system converging on equilibrium. In addition, the substantial growth in the size of PJM created signiﬁcant locational differences in market conditions. Market conditions in the PJM market in 1999 were relatively geographically homogeneous while market conditions in 2007 are quite geographically heterogeneous with respect to both energy and capacity markets. Geographical differences in market conditions are reﬂected in zonal net revenues which in turn reﬂect differentials in locational marginal price (LMP) across the system and illustrate the substantial impact that locational prices have on economic incentives. The level of energy market revenues varied signiﬁcantly by zone, with net revenues higher in the more congested eastern zones than in the less congested western zones. As an example, for a new coal plant, while the PJM average net revenue in 2006 was $182 461 per MW-day, the maximum zonal coal plant net revenue was $259 572 in the PEPCO control zone and

EBL

31

For a detailed description of the calculation of net revenues, see PJM Market Monitoring Unit’s 2006 State of the Market Report.

Competitive Electricity Markets

376

Table 10.1. Total net revenue and 20-year, levelized ﬁxed cost for new entry CT, CC and CP generators CT

1999 2000 2001 2002 2003 2004 2005 2006 Avg

CC

CP

Economic Dispatch Net Revenue

20-Year Levelized Fixed Cost

Economic Dispatch Net Revenue

20-Year Levelized Fixed Cost

Economic Dispatch Net Revenue

20-Year Levelized Fixed Cost

$74,537 $30,946 $63,462 $28,260 $10,565 $8,543 $10,437 $14,948 $30,212

$72,207 $72,207 $72,207 $72,207 $72,207 $72,207 $72,207 $80,315 $73,221

$100,700 $47,592 $86,670 $52,272 $35,591 $35,785 $40,817 $49,529 $56,120

$93,549 $93,549 $93,549 $93,549 $93,549 $93,549 $93,549 $99,230 $94,259

$118,021 $134,563 $129,271 $112,131 $169,510 $133,125 $228,430 $182,461 $150,939

$208,247 $208,247 $208,247 $208,247 $208,247 $208,247 $208,247 $267,792 $215,690

Note: See PJM Market Monitoring Unit’s 2006 State of the Market Report, Vol. I, p.16.

the minimum was $107 531 in the DLCO control zone. The higher LMPs in the eastern PJM zones, reﬂecting transmission limitations and congestion, have a positive impact on the incentive to invest in those areas. While net revenues reﬂect the locational contribution of energy prices, pre-RPM capacity prices are non-locational. While net revenues reﬂect underlying locational supply and demand conditions in the energy market, the contribution of capacity market revenues to net revenues reﬂects only aggregate supply and demand conditions. The result is an attenuated locational investment signal. Prices in the capacity market reﬂected the overall excess supply of capacity in the PJM footprint but did not reﬂect the substantially tighter capacity market conditions in eastern PJM. As an illustration of that fact, the owner of several units in New Jersey notiﬁed PJM of its intent to retire those units, because net revenues to the units were not covering the units’ avoidable costs. PJM’s analysis indicated that the units were needed for reliability. It is a clear indication of market design issues if a unit can be required for reliability, yet the market design does not provide adequate revenues to cover even annual avoidable costs. The PJM tariff provides, in such a case, that the units may be paid the equivalent of rate base, rate-of-return revenues with FERC approval until transmission upgrades are adequate to permit their retirement or the capacity market redesign results in higher locational net revenues. The net revenue performance of the PJM markets over the past eight years leads to the conclusions that capacity markets are needed and that capacity market design matters. Despite the fact that PJM had relatively well-developed energy, ancillary services, and capacity markets over the period, net revenue performance was not consistent with underlying locational conditions in the capacity market. While the pre-RPM capacity market design would have produced better results if PJM had remained in its original physical conﬁguration, the absence of a locational capacity market in the presence of substantial variations in local capacity market conditions was a signiﬁcant market design weakness. Substantial integrations and associated increases in the physical size of the PJM footprint resulted in increased overall excess capacity, associated moderating pressures on energy market prices and on capacity market prices, and signiﬁcant differences in local capacity

EBL

The Evolution of PJM’s Capacity Market

377

market conditions, without an equilibrating mechanism to ensure that local market prices reﬂected local market conditions. 10.6. The RPM Modiﬁcations to PJM Capacity Market Design The essential elements of the RPM modiﬁcations to the PJM capacity market design include an annual market, a forward market, locational capacity markets, scarcity pricing of capacity via a deﬁned demand curve, explicit links to the energy and ancillary services markets, incentives to provide reliability, and clear market power rules including a must offer requirement. The RPM modiﬁcations also include an opt-out provision that permits individual LSEs to not participate in the capacity market and to maintain target levels of reliability in other ways. The RPM design replaced the daily market with an annual market. This is a key design feature because it means that the capacity market clearing price covers an entire year, but only a year. While the volatility of prices is likely to be less for an annual market than a daily market, even the reliance on an annual price has implications for the incentive to build new capacity as the annual price may ﬂuctuate and may result in a relatively volatile future revenue stream for capacity resources. Nonetheless, creation of a stable market design, which creates reasonable expectations of revenue adequacy, is critical. Both volatility and risk are likely to be greater the smaller the locational capacity market size compared to the minimum size of new entry as even a single new entrant can depress the locational price for a signiﬁcant period under some conditions. A preferable design would incorporate a multi-year price for new entry in order to limit volatility and price impacts particularly when locational capacity markets meet criteria related to the size of the local market compared to the minimum size of entry. The RPM design replaces a day-ahead market with a three-year-ahead forward market by running auctions to meet capacity obligations three years in the future. The beneﬁt of a forward auction is that it permits new entry to compete directly with existing units and to set the clearing price when new entry is required to clear the market. Another beneﬁt of a forward auction is that it tends to reduce boom–bust cycles associated with overand then under-investment. Again, a preferable design would permit new entry to set the price for a multi-year period. The RPM design replaces a single capacity market with locational capacity markets in a manner analogous to locational energy markets but with a substantially less granular deﬁnition of locational. While the energy market design accommodates different prices at every bus or node on the system, the RPM design is based on transmission zones, the service territories of the transmission owning utilities in the PJM footprint. PJM deﬁnes the reliability-based requirement for capacity in each transmission zone and the ability to import capacity resources into each zone based on the transmission capability. A zone or set of zones is constrained when the requirement for capacity in the zone can be met only by capacity resources in the deﬁned area, taken out of merit order. While this level of locational pricing is an improvement, it is still not dynamically based on the underlying local supply and demand conditions in areas smaller than zones and could result in an understatement or overstatement of the capacity market conditions in a sub-zone and a corresponding mismatch between the need for capacity in such a location and the price of capacity for that location. The RPM design incorporates scarcity pricing via a deﬁned demand curve, also termed the variable resource requirement in PJM’s ﬁling. The pre-RPM capacity market includes a demand curve which included a ﬁxed demand for capacity and scarcity pricing via

EBL

Competitive Electricity Markets

378

a maximum price equal to a multiple of the CDR. The RPM demand curve shape is based on the net cost of new entry as well as judgments about the appropriate maximum price, the level of capacity at the cost of new entry, and the level of capacity required for reliability. The primary difference between the RPM and pre-RPM demand curves is that the RPM demand curve introduces some slope to the demand curve in the area around the target level of capacity. (See Fig. 10.3.) The slope will reduce price volatility to some extent. The demand curve determines the clearing price when the supply curve directly intersects the demand curve. If there is inadequate supply to intersect the demand curve, the price is determined from the demand curve based on the maximum quantity of available supply. The demand curve can have a uniform shape for the entire PJM footprint or be differentiated by locational differences in the cost of new entry and net revenue. Binding constraints create incremental demand curves for speciﬁc zones or groups of zones, with the same general determinants as the overall demand curve, but with the associated MW determined by incremental capacity requirements in the location. In order for the energy market and capacity markets to equilibrate, net revenues from the energy and ancillary services markets must be reﬂected in the capacity market parameters. When energy and ancillary services markets net revenues are higher, capacity market prices and quantities would be lower and vice versa. The design goal is to produce total market net revenues at a level sufﬁcient to induce entry when it is needed. The alignment in the RPM design is incorporated, in part, in the demand curve parameters via the net cost of new entry and, in part, in the supply curve via the market power mitigation mechanism which deﬁnes the net cost of existing capacity. The RPM demand curve is based, in signiﬁcant part, on the net cost of new entry. (See Fig. 10.3.) The net cost of new entry is deﬁned to be the cost of constructing a new combustion turbine with the lowest cost design, net of the revenues that such a unit would

EBL

$300.00 $275.00

120,862 MW $257.81/MW-Day

A

$250.00

Point 1

$225.00

B

Demand

$/MW-Day

$200.00

Point 2

$175.00

125,238 MW $171.87/MW-Day

$150.00

C

$125.00 $100.00 129,614 MW $34.37/MW-Day

$75.00

Point 3

$50.00 129,614 MW $0.00/MW-Day

$25.00

D

Point 4

Capacity (Unforced MW) Fig. 10.3. 2007–2008 PJM RPM Aggregate demand curve.

130,486

127,279

122,415

120,124

118,191

116,338

113,188

56,406

28,677

13,121

5,510

2,004

526

99

0

$0.00

The Evolution of PJM’s Capacity Market

379

earn. By deﬁnition, the net cost of new entry is an estimate, both with respect to the cost of new entry and the market revenues. In the RPM, the estimate of the gross cost of constructing the new unit was based on actual offers from a ﬁrm that would build a unit for the stated price.32 The estimate of the net market revenues for the new entry unit was based on historical LMPs in PJM, estimates of the fuel and variable operation and maintenance expense to operate the unit, and assumptions about the actual operation of such a unit. The highest price portion (Segment A) of the demand curve (Fig. 10.3) is equal to 1.5 times the net cost of new entry, the price at the next inﬂection point (Point 2) is equal to the net cost of new entry and the price at the next inﬂection point (Point 3) is equal to 0.2 times the net cost of new entry, where each price is converted to an unforced capacity basis. The MW quantity corresponding to the highest price (Point 1) is equal to the overall reliability requirement less the demand that opts out of the RPM, adjusted downward by the ratio of one plus the reserve margin less three per cent to one plus the reserve margin, less any demand side resources adjusted for the demand side resources included in the opt out. The quantity at the next inﬂection point (Point 2) is the overall reliability requirement less the demand that opts out of the RPM, adjusted upward by the ratio of one plus the reserve margin plus one per cent to one plus the reserve margin, less the demand side resources adjustment. The quantity at the next inﬂection point (Point 3) is equal to the overall reliability requirement less the demand that opts out of the RPM, adjusted upward by the ratio of one plus the reserve margin plus ﬁve per cent to one plus the reserve margin less the demand side resources adjustment. The demand curve shown in Fig. 10.3 is for the entire RTO without accounting for locational differences. To the extent that energy market prices are higher, net revenues will be higher, the net cost of new entry will be lower, the demand curve will be lower and capacity prices will be lower. A new feature in the RPM is the updating of the net revenue offset in determining the parameters of the demand curve. The pre-RPM demand curve uses a net cost of entry estimate that has not changed for approximately thirty years. The primary issues in this part of the alignment mechanism are the time period over which the net revenues are calculated and the extent to which the estimate of the gross cost of new entry captures all the actual costs and risks of entry including locational differences. The longer the time period over which the net revenues are calculated, the more attenuated is the net revenue offset and the more likely it is to result either in an underestimate or an overestimate of the actual net cost of entry. RPM uses average net revenue over six years for the ﬁrst three auctions and uses a three-year average thereafter. The alignment between energy and capacity markets in the RPM design is also incorporated in the supply curve via the market power mitigation mechanism which deﬁnes the net cost of existing capacity. Part of the RPM market power mitigation mechanism is the calculation of unit-speciﬁc offer caps which apply to existing units when the overall market or locational market is not structurally competitive. This component of the RPM design is new, as there are no market power mitigation rules in the pre-RPM design. The unit-speciﬁc offer caps are based on unit-speciﬁc avoidable costs, net of actual unit-speciﬁc revenues. These offer caps deﬁne the shape and location of the supply curve. To the extent that actual energy prices received by individual units are higher, capacity offer prices will be lower, the capacity supply curve will be lower, and capacity prices will be lower, all

EBL

32

PJM Interconnection, L.L.C. (2005). Reliability Pricing Model Filing. Docket Nos. ER05-1410-000 and EL05-148-000, 31 August, Tab I, “Afﬁdavit of Ray L Pasteris.”

380

Competitive Electricity Markets

else equal. Again, as with the demand curve, the longer the time period over which the net revenues are calculated, the more attenuated is the net revenue offset and the more likely it is to result either in an underestimate or an overestimate of the actual incremental cost of capacity for existing units. The RPM capacity market design represents a signiﬁcant advance over the current capacity market design because RPM has explicit market power mitigation rules designed to permit competitive, locational capacity prices while limiting the exercise of market power. Given that RPM will increase already endemic structural market power through the creation of smaller, local capacity markets with more concentrated ownership, the market power mitigation rules are necessary to ensure competitive outcomes in the RPM construct. The RPM construct is consistent with the appropriate market design objectives of permitting competitive prices to reﬂect local scarcity conditions (by incorporating locational demand curves) while not relying on the exercise of market power to achieve the design objective. In general, the market power mitigation rules are narrowly targeted to speciﬁc market structures that create the conditions for the potential exercise of market power. The unitspeciﬁc offer caps in the market power mitigation rules apply only where offers from new entrants are not required in order to clear the market. Mitigation is not applied to new entrants, rather competitive forces are relied upon to provide competitive prices when new entry is required, with some broad tests to ensure that market power is not exercised by new entrants. Mitigation based on unit-speciﬁc offer caps is applied only in the situation where the relevant market structure fails the market structure tests and there is enough existing capacity to meet the demand for capacity. The market structure test is the three pivotal supplier test, also used for locational market power mitigation in the energy market.33 Mitigation is applied only if the actual offers exceed the offer cap and if the offer would increase the market clearing price in the absence of mitigation. The proposed mitigation cannot reduce a scarcity price. When existing capacity is not adequate to serve the load in a market, the market clearing price is determined either by new entry and/or by the demand curve. When required, the mitigation of offers from existing units is based on the incremental cost of such capacity, which is the competitive price of existing capacity. The incremental cost of existing capacity equals total annual avoidable costs less net revenue from other PJM markets. The incremental cost of existing capacity also includes the annual costs associated with any new investment in the unit required to maintain its viability as a generating unit. While a capacity market design must incorporate incentives to build new capacity and to continue to invest in existing capacity, it must also incorporate incentives for the units receiving capacity payments to perform and provide energy when it is needed. The actual reliability, which is the goal of capacity markets, is the reliable provision of energy. The RPM design incorporates incentives to provide energy during deﬁned critical periods by using a deﬁnition of the forced outage rate (EFORd) based on performance during those periods. To the extent that a unit does not perform during those periods, the capacity payment to which it is otherwise entitled is reduced, with a minimum payment equal to half of the clearing price for the relevant operating year. Units are excused from the performance incentive if outages are deemed outside management control. This incentive

EBL

33

See PJM Market Monitoring Unit’s 2006 State of the Market Report, Vol. II, Appendix J, “Three Pivotal Supplier Test.” The three pivotal supplier test measures the degree to which the supply from three owners is required in order to meet demand in a market.

The Evolution of PJM’s Capacity Market

381

to perform, like the net revenue offset, is attenuated. A preferable design would include no minimum payment and would have a direct proration of the capacity payment based on hours during which units were needed but did not perform, without exceptions. The RPM design also includes a must offer requirement. Physical withholding is a potentially proﬁtable strategy for exercising market power in the aggregate market or in local markets. The must offer rule requires that market sellers offer all of their PJM capacity resources to the market or they will not be permitted to sell any withheld capacity in any RPM auction. Suppliers may export capacity with a ﬁrm contract or by offering an identiﬁed opportunity cost. While there was a must offer rule in the ﬁrst years of the pre-RPM capacity market, the rule was eliminated in 2001. If the must offer rule does not provide an adequate incentive to offer capacity resources to the market, and if withholding results in a market price increase of ﬁve per cent or greater compared to the price absent withholding, a ﬁling with FERC and a postponement of the ﬁnal clearing of the auction until the issue is resolved are triggered. The RPM design permits LSEs to opt out of the capacity market, but they are required to meet deﬁned capacity obligations. Such LSEs are responsible for obtaining capacity and meeting capacity requirements but their incentives to obtain capacity are not based on the market. (The most likely alternative is a state public utility commission regulatory incentive to build capacity.) It would be preferable not to have an opt-out provision from this critical PJM market, as one of the beneﬁts of the RTO approach is the creation of an overall market design that permits all members to participate in deep and liquid markets governed by the same rules. The RPM capacity market, by design, is tightly linked to the energy market. Permitting an opt-out from the capacity market creates a potential risk to the overall design. Investments in transmission that increase the transfer capability into a locational capacity market may participate in the RPM auction process. Such transmission investments are paid based on the difference between the locational RPM clearing prices in the source and sink locations. In general, however, transmission investments are currently compensated under a traditional regulatory model. The incentives facing large, vertically integrated transmission owners are complex. The fact that a transmission owner could invest in transmission capability that changes locational capacity prices creates a risk for investors in new generation. The risk facing a merchant generation unit, which is built based on expectations about high locational capacity prices, is that the locational price difference may be eliminated or reduced as a result of transmission construction by an integrated utility facing non-market incentives. The PJM process for determining whether to require “economic” transmission investments in addition to reliability related transmission investments depends on an evaluation of costs and beneﬁts using multiple tests without an identiﬁed decision rule. The determination of transfer capability into transmission zones in the RPM market is also somewhat opaque to market participants. These non-market factors make the expected outcomes less predictable. While investors in new generation face both market and non-market risks, a clear set of market rules governing transmission investment would make such investment more predictable and reduce the risk facing investors in new generation.

EBL

10.7. Conclusions Wholesale electric power markets are affected in signiﬁcant ways by externally imposed reliability requirements. A regulatory authority administratively determines the acceptable level of reliability, which is enforced through a requirement to maintain a target level of installed or unforced capacity. The requirement to maintain a target level of capacity

382

Competitive Electricity Markets

can be implemented via a variety of market mechanisms, including capacity markets and scarcity pricing designs and non-market mechanisms that may include state public utility commission mandates and incentives to construct capacity and government construction of generation. Regardless of the mechanism, the exogenous requirement to construct capacity in excess of what is constructed in response to energy market signals has an impact on energy markets. The reliability requirement results in maintaining a level of capacity in excess, on average, of the level that would result from the operation of an energy market alone. The result of that additional capacity is to reduce the level and volatility of energy market prices and to reduce the duration of high energy market prices. This, in turn, reduces net revenue to generation owners which reduces the incentive to invest. The application of reliability requirements means that scarcity conditions in the energy market occur with reduced frequency. Enforced levels of reliability require units that are only directly used and priced under relatively unusual load conditions. Thus, the energy market alone frequently does not directly value the resources needed to provide for reliability, although the contribution of the energy market will be more consistent with reliability signals if the energy market appropriately provides for scarcity pricing when scarcity does occur. The exogenous reliability requirements exist and must be met whether capacity markets are incorporated or whether the market design relies solely on energy markets. While it is straightforward to incorporate the required level of capacity in a capacity market design, that is not the case for an energy-only market design. If an energy-only market is to achieve an administratively determined level of reliability, energy prices must be managed so as to achieve that level of reliability. There is no market-based, self-correcting signal that would lead an energy market to such a level of reliability.34 Demand for energy is quite inelastic and the physical capability does not yet exist to permit individual customers to signal or to receive their desired level of reliability. Managing reliability in an energy-only market means managing net revenues, just as it does in a market design which includes capacity markets, which, in turn, means managing scarcity pricing. Managing scarcity pricing so as to produce a deﬁned level of reliability means taking account of the relationship between actual levels of demand and supply in real time and administratively establishing prices that reﬂect deﬁned levels of scarcity, that produce desired levels of net revenue, and that produce the associated investment incentives. Managing scarcity pricing is also required to ensure that market power does not result in inefﬁcient wealth transfers without corresponding investment incentive effects. While increasing the frequency and duration of appropriate scarcity pricing is straightforward, doing so in a way that can be expected, by itself, to result in the desired level of reliability is not. While capacity market designs are complex, the successful design of such a scarcity pricing regime to achieve a target level of capacity in an energy-only market design is likely to be at least as complex.35 Even with a capacity market, energy market design must permit scarcity pricing, when such pricing is consistent with market conditions and constrained by reasonable rules to ensure that market power is not exercised. Scarcity pricing is also part of an appropriate incentive structure facing both load and generation owners in a working wholesale

EBL

34 Cramton, and Stoft, The convergence of market designs for adequate enerating capacity with special attention to the CAISO’s resource adequacy problem.; and Joskow, and Tirole, Reliability and Competitive Electricity Markets. RAND J. Econ. 35 See, for example, Hogan, W.W. (2005). On an ‘energy-only’ electricity market design for resource adequacy. 23 September.

The Evolution of PJM’s Capacity Market

383

electric power market design. Scarcity pricing must be designed to ensure that market prices reﬂect actual market conditions, that scarcity pricing occurs in well-deﬁned stages with transparent triggers and prices, and that there are strong incentives for competitive behavior and strong disincentives to exercise market power. Such administrative scarcity pricing is a key link between energy and capacity markets. With a capacity market design that incorporates scarcity rents in the energy market, scarcity pricing can be a mechanism to increase reliance on the energy market as a source of revenues and incentives in a competitive market without sole reliance on the energy market to signal the need for the target level of capacity. Net revenue in PJM has been below the level required to cover the full costs of new generation investment on average for all unit types for the entire eight-year PJM market period. The issue is how to understand this phenomenon and how to address it within the context of competitive markets. The level of net revenues in PJM markets is not the result of the $1000 per MWh offer cap, of local market power mitigation, or of a basic incompatibility between wholesale electricity markets and competition. Competitive markets can, and do, signal scarcity and surplus conditions through market clearing prices. The net revenue experience of PJM has been the result of the interaction of several factors including weather-driven demand, the shape and location of the energy supply curve, the level of excess capacity and the market design. The nature of the supply curves for energy and capacity has been shaped by the integrations that occurred in PJM between 2002 and 2005. The PJM markets, as a result, cannot be said to reﬂect an endogenously determined equilibrium. The fact that investors’ expectations have not been realized in every year does reﬂect aggregate supply–demand fundamentals in PJM markets as they were designed, but those fundamentals have been, in signiﬁcant part, the result of exogenous factors. However, the capacity market design did not permit the market signals associated with local market fundamentals and therefore the net revenue did not reﬂect those local fundamentals. The actual performance of PJM’s markets between 1999 and 2006 supports the need for a capacity market and for a capacity market with improved design features. Even with the pre-RPM capacity market, there were signiﬁcant revenue shortfalls. An improved capacity market design and improved scarcity pricing are required in order to ensure that revenue adequacy exists, that investment incentives exist, that performance incentives exist, and that ultimately the market is competitive and self-sustaining and provides power at the lowest possible cost. The observed revenue shortfalls would have been larger without the existing capacity markets, although the difference was relatively small in recent years. An energy-only market design would have been unlikely to have produced revenue adequacy during this period even with a much higher aggregate market offer cap. There were relatively few periods of scarcity in the eight years. Reliance on extremely high energy prices and revenues during a few hours that occur unpredictably and sometimes only every several years would create a volatile and uncertain revenue stream. The investor expectations and associated investment behavior in such an environment are likely to reﬂect that volatility in the form of a risk premium. While there have been revenue adequacy issues in the PJM market design with a capacity market, it is important to point out that there is no wholesale power market in the United States that has successfully relied on an energy-only market to provide adequate capacity.36

EBL

36

See Chapter 9 by Adib, Shubert, and Oren in this volume for a discussion of energy-only market designs.

384

Competitive Electricity Markets

While RPM can be expected to improve overall market performance, RPM still needs signiﬁcant improvement. Areas for improvement in the RPM design include more granular locational prices, a multi-year price for new entry to reduce volatility, a stronger link between net revenues from the energy and ancillary services markets and prices in the capacity market, and a stronger link between the actual performance of capacity resources and the prices paid for capacity. A combination of these improvements in the RPM design and increased reliance on appropriate scarcity pricing in the energy market will provide more efﬁcient performance signals and reduce reliance on revenues from the capacity market. While this will not ultimately mean the demise of capacity markets, it will result in strengthened energy market price signals which are more locationally accurate than capacity market prices. This, in turn, will provide better price signals for investment in capacity and for investment in demand side resources that can respond to such signals. Capacity markets are not a panacea, but a properly designed capacity market plays a critical, if circumscribed, role in wholesale power market design. Bibliography Key FERC Orders re Competitive Electricity Markets Order No. 888 (1996). Transmission Open Access. Promoting Wholesale Competition Through Open Access Non-discriminatory Transmission Services by Public Utilities; Recovery of Stranded Costs by Public Utilities and Transmitting Utilities, 24 April. Order No. 888-A (1997). Reafﬁrming and Clarifying Terms of Order 888 in Regards to Open Access Transmission Services and Recovery of Stranded Costs, 4 March. Order No. 888-B (1997). Order Afﬁrming, With Certain Clariﬁcations, The Fundamental Calls Made in Order 888-A in Regard to Promoting Wholesale Competition Through Open Access Nondiscriminatory Services etc., 25 November. Order No. 889-B (1997). Final Order: Order Denying Rehearing of Order 889-A in Regards to Open Access Same-Time Information System and Standards of Conduct (OASIS), 25 November. Order No. 888-C (1998). Order on Rehearing, Denying Otter Power Co. Rehearing Request, 20 January. Order No. 2000 (1999). Final Rule, Regional Transmission Organizations (RTO), 20 December. Order No. 2000A (2000). Regional Transmission Organizations (RTO), Order on Rehearing, 25 February. Docket Nos. EL01-118-000, EL01-118-001 (2003). Market Behavior Rules, 17 November. Docket No. EL01-118-003 (2004). Order Clarifying Market Behavior Rules, 19 May. Docket No. PL05-1-000 (2005). Policy Statement on Market Monitoring Units, 27 May. Order No. 667 (2005). Repeal of the Public Utility Holding Company Act of 1935 and Enactment of the Public Utility Holding Company Act of 2005 (Final Rule), 8 December. Order No. 667-A (2006). Repeal of the Public Utility Holding Company Act of 1935 and Enactment of the Public Utility Holding Company Act of 2005 (Final Rule), 24 April. Order No. 681 (2006). Long-Term Firm Transmission Rights in Organized Electricity Markets (Final Rule), 20 July. Order No. 681-A (2006). Long-Term Firm Transmission Rights in Organized Electricity Markets (Order on Rehearing), 16 November. Order No. 890 (2007). Preventing Undue Discrimination and Preference in Transmission Service (Final Rule), 16 February.

EBL

Key FERC Orders re PJM Electricity Markets Docket Nos. OA97-261-000, ER97-1082-000 (1997). Order Accepting For Filing and Suspending Proposed Pool-Wide and Single-System Holding Company Open Access Transmission Tariffs and Revised Tariffs, and Deferring Further Action, 28 February.

The Evolution of PJM’s Capacity Market

385

Docket Nos. OA97-261-000 et al. (1997). Order Conditionally Accepting Open Access Transmission Tariff and Power Pool Agreements, Conditionally Authorizing Establishment of an Independent System Operator and Disposition of Control over Jurisdictional Facilities and Denying Rehearings, 25 November. Docket No. ER99-196-000 (1999). Order Accepting, as Revised, PJM Capacity Credit Markets, 13 January. Docket No. ER97-3729-000 (1999). Order Approving PJM Supporting Companies’ Request for MarketBased Pricing Authority, 10 March. Docket No. ER98-3527-000 (1999). Order Approving Market Monitoring Plan as Modiﬁed, 10 March. Docket No. ER99-2028-000 (1999). Order Conditionally Accepting Compliance Filing, 13 April (Tariff revisions for FTR Auctions). Docket No. ER00-1630-000 (2000). Letter Order accepting amendments to the Open Access Transmission Tariff and Operating Agreement providing for market-based pricing and new market rules for regulation service, 12 April. Docket No. ER00-1849-000 (2000). Letter Order accepting revisions to the Open Access Transmission Tariff and Operating Agreement, containing a modiﬁed two-settlement system that incorporates increment and decrement bids into the day-ahead energy market, 18 May. Docket No. ER00-3090-000 (2000). Order Accepting and Suspending Filing, 26 July (Load Reduction Pilot Program). Docket No. ER01-1671-000 (2001). Order Accepting Tariff Sheets as Modiﬁed, 30 May (Emergency and Economic Load Response Programs). Docket No. RT01-2-000 (2001). Order Provisionally Granting RTO Status, 12 July. Docket Nos. ER02-2519-000, ER02-2519-001, ER02-2519-002 (2002). Order Accepting Spinning Reserve Market, 31 October. Docket Nos. RT01-2-001, RT01-2-002 (2002). Order Granting PJM RTO Status, Granting In Part and Denying In Part Requests For Rehearing, Accepting And Directing Compliance Filing, and Denying Motion For Stay, 20 December. Docket No. ER03-406-000 (2003). Order Accepting For Filing Proposed Tariff Changes, As Modiﬁed, 12 March (Annual FTR auction). Docket Nos. ER05-1410-000 and EL05-148-000 (2006). Order Denying Rehearing and Approving Settlement Subject To Conditions, 22 December (RPM Capacity Settlement Agreement).

EBL

Additional PJM Resources Market Monitoring Unit, PJM Interconnection. 1999 State of the Market Report (June 2000). Market Monitoring Unit, PJM Interconnection. 2000 State of the Market Report (June 2001). Market Monitoring Unit, PJM Interconnection. 2001 State of the Market Report (June 2002). Market Monitoring Unit, PJM Interconnection. 2002 State of the Market Report (5 March 2003). Market Monitoring Unit, PJM Interconnection. 2003 State of the Market Report (4 March 2004). Market Monitoring Unit, PJM Interconnection. 2004 State of the Market Report (8 March 2005). Market Monitoring Unit, PJM Interconnection. 2005 State of the Market Report (8 March 2006). Market Monitoring Unit, PJM Interconnection. 2006 State of the Market Report (8 March 2007). Additional analytical reports produced by the Market Monitoring Unit can be found at http://www.pjm.com/markets/market-monitor/reports.html. Additional resources about PJM Markets can be found at www.pjm.com.

Additional electricity market resources Crampton, Peter and Stoft, S. (2005). A capacity market that makes sense. Elec. J., 18 (August/September), 43–54. Cramton, P. and Stoft, S. (2006). The convergence of market designs for adequate generating capacity with special attention to the CAISO’s resource adequacy problem. April. Available at: http://stoft.com/.

386

Competitive Electricity Markets

Hobbs, B.F., Iñón, J., and Stoft, S.E. (2001). Installed capacity requirements and price caps: Oil on the water, or fuel on the ﬁre? Elec. J., 14 (July), 23–34. Hogan, W.W. (2005). On an ‘energy-only’ electricity market design for resource adequacy. 23 September. Available at: http://www.whogan.com/. Hogan, W.W. (1998). Competitive Electricity Market Design: A Wholesale Primer. John F. Kennedy School of Government, Harvard University, Cambridge, MA, 17 December. Available at: http://www.whogan.com/. Hogan, W.W. (1992). Contract Networks for Electric Power Transmission, revised. John F. Kennedy School of Government, Harvard University, Cambridge, MA, February. Available at: http://www.whogan.com/. Hogan, W.W. and Ring, B.J. (2003). On Minimum-Uplift Pricing for Electricity Markets, John F. Kennedy School of Government, Harvard University, Cambridge, MA, 19 March. Available at: http://www.whogan.com/ Joskow, P. and Tirole, J. (2007). Reliability and competitive electricity markets. RAND J. Econ. Spring, Vol. 38–1, pp. 60–84. Available at: http://econ-www.mit.edu/faculty/index.htm?prof_id=pjoskow. Kahn, E.P., Cramton, P., Porter, R., and Tabors R. (2001). Pricing in the California Power Exchange Electricity Market: Should California Switch from Uniform Pricing to Pay-as-Bid Pricing? Blue Ribbon Panel Report, Commissioned by the California Power Exchange, 23 January. Available at: http://www.ksg.harvard.edu/hepg/Papers.htm. Lambert, J.D. (2001). Creating Competitive Power Markets: The PJM Model. Tulsa, OK: PennWell Corporation. Schweppe, F.C., Caramanis, M.C., Tabors, R.D. and Bohn, R.E. (1988). Spot Pricing of Electricity. Boston: Kluwer Academic. Stoft, S. (2002). Power System Economics. Piscataway, NJ: IEEE Press.

EBL

Chapter 11 Resource Adequacy and Efﬁcient Infrastructure Investment ALAN MORAN AND BEN SKINNER Institute of Public Affairs, Australia; TRUenergy, Melbourne, Australia

Summary Among the many issues with which electricity market designers have wrestled is how to ensure reliable and uninterrupted supply. The concerns cover both short-run operations and longer-term investment adequacy, the issue on which this chapter is focused. Electricity is jointly supplied to the whole community, has virtually no storage capabilities, and faces a peaky demand with little capacity or desire of consumers to respond to excess demand (and hence price surges) by reducing their demand. In addition, it is subject to political oversight of prices and many facets of supply. Many have argued that as a result there will be “missing money” in the market and that we must, therefore, have a dual market for electricity generation, covering energy on the one hand and capacity on the other. Similar issues are present with electricity transmission where virtually all markets involve forms of regulated supply. This chapter ﬁnds that a reasonably efﬁcient market has been achieved in Australia without regulation of generation. The outcome, which is not without some fragilities, has been due to generally less government intervention than seen in some other markets, with a higher reserve intervention price, less distortive consumer price caps, and a genuine level of retail competition that provides good market signals for new capacity.

EBL

11.1. Introduction A key debate surrounding electricity markets remains the general question of resource adequacy – i.e., can we leave investment in long-lead time and long-lived assets, producing a product essential to every other part of the economy to the chaos of a free market? Many commentators, indeed many market designs, promote capacity obligations to underpin a certain amount of generation investment regardless of energy price signals. These designs, in turn, provoke great debates as to whether they are themselves efﬁcient, or are achieving their objectives, i.e. whether the new investment is sufﬁcient. The notion that energy-only markets cannot provide adequate reliability is most directly addressed within a robust theoretical framework by Stoft (2002), though it derives from 387

388

Competitive Electricity Markets

the concept, developed by Boiteux (1949), of electrical energy being two goods: reliability and immediate power. Oren (2000) has also supported a form of capacity payment as well as an energy market, though as a second best approach in light of a seeming inevitability that governments would always intervene in this market to prevent very high prices. Within Australia, Simshauser (2006) has been an active proponent of capacity markets. Caramanis (1982) was an early advocate of an energy-only market and demonstrated the conditions under which this could operate. He and others looked to the removal of government regulations on pricing and plant development to ensure adequate investment in new capacity. Cramton and Stoft (2006) cite Joskow (2006) in deﬁning the conditions that prevent markets from operating to provide optimal capacity when it is needed. Joskow says: The problems include: [1] price caps on energy …[2] market power mitigation mechanisms that do not allow prices to rise high enough during conditions when generating capacity is fully utilized …[3] actions by system operators that have the effect of keeping prices from rising fast enough and high enough to reﬂect the value of lost load …[4] reliability actions taken by system operators that rely on Out of Market (OOM) calls on generators that pay some generators premium prices but depress the market prices paid to other suppliers, …[5] payments by system operators to keep inefﬁcient generators in service due to transmission and related constraints rather than allowing them to be retired or be mothballed, …[6] regulated generators operating within a competitive market that have poor incentives to make efﬁcient retirement decisions, depressing market prices for energy. All of these problems represent market corruption by the regulatory authorities. In essence, all of them are measures taken to avoid having price undertake its conventional role of determining what is to be supplied to the market. They represent either a mistrust that price will offer the correct signals or that allowing the necessary prices to be visible will spark political concerns. Reviewing the UK market which has had experience of both a capacity payment and the current NETA energy-only market, Roques et al. (2005) are neutral between the two. They argue, however, that the current UK balancing mechanism which has two prices (unlike Australia’s single pool price) mutes signals and should be changed if an energy-only market is to operate effectively. This chapter examines the concerns about resource adequacy in the context of the “energy-only” Australian National Electricity Market (NEM). It argues that the NEM has worked well. Prices have remained among the lowest in the world, reliability has been maintained, and the market has produced new generation investment of the magnitude, type, and timing that has been appropriate. These results point to the superiority, at least in the Australian context, of an energy-only market approach that operates without the potential distortions that separate capacity payments bring. Key reasons for this success include a relatively high wholesale reserve price at $A 10 000/MWh. (Hereafter all $ prices in regard to Australian supplies refer to Australian dollars; as at November 2007 the $A was worth 88 US cents). In the US, price caps are set at much lower levels of $400 in California and $1000 in New England, Midwest, New York, PJM, and Southwest. ERCOT is at $1500 and scheduled to increase to $3000 in 2009. There are other features that have contributed to the NEM’s energy-only market success. These include a relatively unfettered retail market that has allowed robust retail competition which provides appropriate market signals. In addition, the market design includes a transparent and ﬂexible bidding system, including the integration of offers for frequency control and other ancillary services with energy market bids. The bidding rules

EBL

Resource Adequacy and Efﬁcient Infrastructure Investment

389

allow multiple- and short-timeframe changes. Although there are certain constraints on generators’ actions in this regard, these are designed to prevent a generator bidding erratically to destabilize the market and impose costs on competitors. This is further discussed in Section 11.4.3. Even though the Australian market is one of the most lightly regulated in the world, it has its fragilities. These stem from actual or potential government intervention. They include: •

Will provisions for intervention when short-term supply is judged to be inadequate result in a dual market, depress some prices and deter new investment in capacity? • Are all government generation investments genuinely commercial and, if not, will this deter new private investment and reduce capacity by more than is created? • How are we to cope with greenhouse issues which present a risk and some reality of carbon tax/trading schemes? These matters are discussed in the context of the NEM, its history market structure, and outcomes in terms of prices, supply productivity, and reliability. Also addressed are the more intractable problems that seem to be present in ensuring adequate investment in transmission in view of its features as both a competitor and a vehicle for generation. The chapter explores measures to facilitate efﬁcient transmission investment without central planning. 11.2. Market History and Outcomes

EBL

11.2.1. Size and nature of the Australian reticulated energy market The Australian National Electricity Market (NEM), now covering all jurisdictions apart from Western Australia and the Northern Territory and close to 95% of consumers, has been in operation since the late 1990s. It is a market that has some government and regulatory intervention: much of the industry remains in government ownership; some retail caps continue in place; there is regulatory uncertainty regarding environmental conditions attached to new generation plant; and there are seemingly endemic debates that precede new transmission developments. Electricity dominates reticulated energy supply, though gas is also important both in its own right and as a fuel for electricity (gas accounts for about 8 per cent of generation). Figure 11.1 shows the market proﬁle of Australian jurisdictions. 11.2.2. The reforms of the 1990s Historically, Australian electricity supply, like that of most European countries, was reserved for government ownership. This grew up partly because of concerns about natural monopoly that under private enterprise might exploit customers, partly because electricity (and gas) was seen as part of the “commanding heights” of commerce that only government should control. In addition, production and supply of electricity was considered to require a level of coordination that many in politics thought it impossible for competing producers to accommodate. In 1992, Australia’s electricity industry comprised seven jurisdictionally based integrated utilities that had total control over generation and sales within their respective states. Competition from other suppliers and retailers was illegal.

Competitive Electricity Markets

Consumption – GWh equivalent

390

100,000

LPG Natural gas Electricity

90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000

N T

TA S

SA

A W

LD Q

VI C

N SW

–

(LPG = Bottled Propane Gas) Fig. 11.1. Australian energy market – Consumption by state. Note: Gas consumption has been converted from PJ to GWh.

As in a great many countries, the early 1990s saw an increased awareness in Australia of the shortcomings of the integrated electricity industry’s efﬁciency. A better appreciation developed of the nature of the industry. This included a realization that the industry need not operate as an unitary monopoly, and that some considerable economies were being delivered in the England and Wales electricity market, the previously integrated nature of which had been the blueprint for the separate Australian systems. On top of this, private ownership was being recognized as providing efﬁciency premiums over governmentowned systems, not only in the newly privatized England and Wales industry but also in the mainly private systems that had long been standard in the US. Formal reports by government and private economic policy institutions [e.g. Industry Commission (1991); Institute of Public Affairs (1991)] lent weight to the evidence of inefﬁciency in Australia compared to elsewhere. There was also a rare level of political consensus developing in favor of greater competition as a means of improving Australian economic outcomes. A major report (National Competition Policy, 1993) had led to the agreement by the federal government to provide additional funding of the state governments on condition that the latter structurally separated the parts of their network industries that were natural monopolies from those where competition was possible. This was to be followed by the opening up of their local markets to competition. Electricity was the industry where these conditions were most obviously present and was singled out for particular attention. Unbundling the monopolies meant dividing each of the single state government generation and retail businesses into rival ﬁrms. It also meant requiring transmission systems to be opened on the basis of non-discriminatory access and with generators being scheduled on the basis of their bid offers. An important factor in the evolution of the industry into a competitive market was the parlous nature of state government ﬁnances in Victoria and South Australia after a period of barely restrained expenditure increases. In Victoria, the consequent level of debt provided an incoming Liberal (conservative) government with a justiﬁcation for pursuing privatization, which is never a politically popular course in Australia. The

EBL

Resource Adequacy and Efﬁcient Infrastructure Investment

391

Victorian government’s most valuable asset capable of being privatized was the electricity industry. In privatizing the electricity industry, the UK model provided a guide. In advance of the federal government’s requirements to do so, the state government ﬁrst disaggregated the electricity monopoly to bring about structural separation of the generation, transmission, and retail/distribution functions, and to ensure multiple competitive providers for generation and retailing (which was left with distribution but with a clear administrative separation). The natural monopoly poles and wires businesses were regulated under a UK-style price setting regime. 11.2.3. Australia’s market design The Australian National1 Electricity Market was guided by, but also avoided some of the mistakes of, the UK’s original gross pool design. Like the UK, it beneﬁted from historic government ownership by allowing a step-wise transformation without excessive compromise to protect legacy positions. Unlike the UK, however, the federal system of government, with states responsible for energy, presented signiﬁcant challenges. Nevertheless, thanks to a rare alignment of state desires and federal threats and funding, the NEM did form in 1998. The key starting advantages over the UK centralized mandatory pool were: • •

A competitive generator ownership structure; A purist “energy-only” market design, without capacity payments and with selfcommitment without uplift compensation; • A degree of locational pricing through market zones or “regions” without constrained uplift payments; • A ﬁve-minute, “real-time” price, with re-bidding allowed up to the point of dispatch; and • Transmission planning and operation separated from the independent market/system operator.

EBL

Other key features of the NEM included: •

the separation of monopoly networks from retail and transmission with the networks operations (and most augmentations) funded by regulated charges on customers; • a phased introduction of retail competition and associated vesting contracts; and • a “VoLL,” or wholesale price cap, approaching the true cost of consumer interruption. The “pure” nature of the energy price, i.e. unadulterated by forms of uplift, has ensured that generators and retailers trade an identical commodity, and can easily deal in the forward market (see Fig. 11.2). It leaves each of the various players – retailers, customers, and generators – with their own responsibility of ensuring their ongoing viability and proﬁtability. The NEM’s forward markets have achieved quite reasonable turnover and liquidity considering the small physical size, challenging claims that a gross pool design limits forward market participation. Indeed, the Australian market, though based on a 1 Note that the NEM covers all the interconnected Australian grid. Western Australia and Northern Territory are not part of the NEM due to their remote location.

Competitive Electricity Markets

392 Forward market

Spot market

Grid

Wholesale

Generator Hedge $ Trader

Pool $

Futures Exchange

Transmission grid

MSO

Hedge $

Pool $ Retailer

Retail

Connection charges $

Network charges $

Distribution network

Bill $

End use customer

Fig. 11.2. Market design.

pool and spot price, is fundamentally one of contracts, which are settled on a contractsfor-difference basis. Whilst the advantages of locational pricing and energy-only markets are well discussed in the literature, some other less well-known features of the NEM’s dispatch and pricing process have equally contributed to its success. The ﬁve-minute pricing and dispatch cycle allowed effectively real-time balancing of supply and demand, with prices non-ﬁrm until the moment of dispatch. This permits simpliﬁcation of electricity price to one value, the energy price, to which the supplier and consumer are equally exposed.

EBL

•

The market/system operator makes no inter-temporal dispatch decisions. There is no day-ahead pricing nor central commitment. Thus, there is no market exposure to MSO forecasting error and the commensurate uplift charge. • “Ancillary services” are limited to balancing the market within a ﬁve-minute dispatch cycle, where generators and interruptible customers are paid for the service of providing some contingency spinning reserve to control frequency until the next cycle. The power system only needs a small volume of reserve for ﬁve-minute balancing, about 1.5% of underlying demand, and there are many competitive providers. Turnover in that market is about 1% of that in the energy market. These services are largely supplied on a spot bidding system (which is pragmatically linked with the general energy bidding) funded by a separate “causer pays” charge. Even ancillary services have common market prices that can be hedged, although the low and stable price has brought little demand for this. Rather than having various forms of central decisions and administered payments to maintain an orderly power system – for example, centrally guaranteed day-ahead pricing for demand-side response and slow-start committing units – the Australian system leaves the responsibility for taking these decisions to providers who do so in the light of their

Resource Adequacy and Efﬁcient Infrastructure Investment

393

own capabilities and commercial options. If, for operational reasons, they physically need to lock in decisions 24 hours ahead, they can do that contractually and this does not need to be underwritten centrally. There are many problems in Market Operators making forward decisions, including complexity, gaming opportunities, perverse incentives against ﬂexibility, etc., but the most obvious problem is error. Australian electricity demand is notoriously unpredictable, a day-ahead Market Operator who predicts a high demand, will set high day-ahead prices and be embarrassed when the demand fails to eventuate. Moreover, the artiﬁcially high prices will discourage demand that is clearly suppliable. The move to real-time self-commitment was met with skepticism by those who prefer others to take the forecasting risk. But after the decision was made in the mid-1990s, the commercial rewards available by ensuring physical ﬂexibility and speed brought entrepreneurial reactions. Peaking units that for decades had demanded from the operator a minimum of ﬁve hours notice of recall discovered ways to start within two. The same is also true for the relatively insigniﬁcant suppliers from the demand-side. And such improvements in ﬂexibility, in addition to rewarding the supplier, also provide a cost saving bonus to the consumer in general by putting downward pressure on prices. The success of the NEM, notwithstanding it being clouded by a less than minimal set of interventions by governments, appears to corroborate initial analyses that the electricity market is not markedly different from other markets. To be sure, there are externalities, and a failure by one party can have repercussions across a great many others, but this is also true of many other markets with independent agents in the supply chain. And if the instantaneous nature of electricity is unique, other industries’ supply characteristics are converging toward this as modern production methods are characterized by considerable economizing in inventories and other buffers.

EBL

11.2.4. Market structure The history of state government-owned monopoly allowed governments to design a structure alongside a market. This implemented the mid-1990s prevailing view of an ideal industry structure, with numerous generators, stapled retailer/distributors, and large monopoly transmission companies. Since that time, the notable new trends are: • •

Self-imposed separation of network and retail businesses; Aggregation of network businesses, including distribution and transmission, with regulatory blessing; • Aggregation of retailing without much regulatory acceptance; and • Vertical integration of generation and retailing despite regulatory resistance. 11.2.4.1. Generation Despite numerous ownership transactions, the generation sector remains about as aggregated as it was when ﬁrst split by the governments (see Fig. 11.3). In a national sense, there is a highly competitive market in terms of capacity, though at particular times a supplier can ﬁnd itself with market power. The market is more concentrated when viewed in a locational sense. In particular, Tasmania is dominated by one government-owned generator, and South Australia’s largest generator comprises one power station that has about 40% of local capacity. Although there is a fairly robust interconnection capacity, supply between the regions is neither inﬁnite nor risk-free. But the almost limitless opportunity

Competitive Electricity Markets

394

M

O ell pt im Ec a og Fl en in de rs O th er

IP

w an St

ac qu

ar ie Er ar D in e l g en ta er gy Ya llo Sno ur w y n ( C CL S P) en e Lo rgy y ya n Ta g ro ng N R G

16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0%

Fig. 11.3. Generation ownership: Capacity by market share. Source: ESAA (Energy Supply Association of Australia) (2006) Electricity Gas Australia 2006.

for new power stations in the NEM presents a very real and effective new-entry threat to most locations. Whilst the governments created a generally competitive generator structure, they faced union resistance against privatizing. Only Victoria and South Australia were able to consummate the process. Queensland and NSW have held their generators on a “forthcoming auction” footing for a decade. This makes planning for the generators themselves difﬁcult, whilst also presenting a sovereign risk for private investors who fear that the governmentowned competitors may act non-commercially.

EBL

11.2.4.2. Retail businesses The original retail franchisees were initially stapled with ring-fenced regulated monopoly distribution businesses over the same geographic area. This was initially feared to be a barrier to retail competition, but it subsequently became irrelevant as private owners realized that the two activities were very different, and chose to specialize by de-merging. This has also occurred with the state based retailer/distributors, informally in the case of the largest one, NSW’s EnergyAustralia, which had a retail alliance with the private generator/retailer International Power, and formally for the Queensland businesses, the retail arms of which have been privatized. Others, 2% Aurora, 3% Country, 8%

AGL/Actew, 24%

Integral, 9% Synergy, 9% TRUenergy, 7%

Origin, 22%

Energy Australia, 17% Fig. 11.4. NEM major electricity retail market shares (by customer numbers). Source: UBS (Union Bank of Switzerland), 2006.

Resource Adequacy and Efﬁcient Infrastructure Investment

395

At the same time, there was some retail aggregation (by government decision in NSW and driven by commercial pressures in Victoria). This reﬂected a view that the originally estimated minimum competitive size of around 0.5 m customers for major retailers was too low. Despite that, many niche retailers with far fewer customers have proﬁtably entered. The market shares of major retailers are illustrated in Fig. 11.4. The big retailers have tried some further tactics: going “dual fuel,” selling electricity and natural gas and, more controversially, merging with generators to form vertically integrated energy businesses. This was challenged by the competition regulator as limiting market entry into either generation or retailing; however, the regulator’s position was overturned in court. Most retailing is now or will shortly be vertically integrated in some form with generation, yet by all measures, competition continues to strengthen at each end. 11.2.4.3. Distribution businesses Specialist regulated infrastructure owners began to accumulate network businesses. As they are not by deﬁnition exposed to competition, the regulator has had no objection. However, the business models of the enlarged ﬁrms have proven especially challenging for price regulators to monitor efﬁcient costs. Economic regulation will need to either become more intrusive – such as the US model – or transform to another model entirely where actual costs are less relevant. 11.2.4.4. Transmission and market system operator (MSO) As a residue of the state-based system, the NEM has ﬁve transmission providers: one per state. This is clearly inefﬁcient and state-owned transmission systems are often criticized for using their inﬂuence to favor intra-state over national solutions to transmission construction. The states did agree to combine the market/system operator to one company, the National Electricity Market Management Co (NEMMCO) that operates all the transmission systems and generators.

EBL

11.2.4.5. Developments in market structure Deregulation having shaken the ossiﬁed system up, we are now seeing the pieces reassembling. There is certainly the move toward retailing and generation forming alliances and cross ownerships. This reﬂects the importance of risk minimization, especially since the price cap is set at a relatively high $10 000/MWh. Even so, there is no move toward a full integration and few consider this to be likely – in this respect, something similar to the oil industry is taking place with ﬁrms adopting a spectrum of supply acquisition ranging from spot to ownership. At the same time we are seeing a voluntary divorce, which nobody envisaged, between distribution and its formerly linked retailing activities – this is also happening, in a somewhat surreptitious way, with the state-owned outﬁts, which are also forming marketing alliances with generation. This is driven by risks and synergies. The fact that retailers also own some generation does not undermine the market since, even without any requirements for Chinese walls, retail buyers would not favor their afﬁliate. To do so would jeopardize their abilities to contract with non-afﬁliates and would thereby undermine their abilities to perform a key function – risk management. In this respect, there is an analogy with the motor industry where assemblers buy components from each other, including for new models, but the component suppliers would not reveal conﬁdential information to afﬁliates, because if they did so they would lose all third-party business.

Competitive Electricity Markets

396 11.3. Market Outcomes

11.3.1. General experiences in productivity In the US, there is little evidence of private ownership and other divestment, bringing about increased efﬁciency. Bushnell and Wolfram (2005) estimate at best a 2% improvement in fuel efﬁciency. Others examining industries that were previously largely government-owned – for example, Newbery and Pollitt (1997) – ﬁnd considerable gains with respect to privatizations in England and Wales. Similarly Fabrizio et al. (2004) found, “The performance gain of an IOU plant in a restructured regime relative to MUNI plants over the same period is …on the order of 15% reductions in employees and 20% reductions in nonfuel expenses.” Australia’s experiences show improvements across a range of indicators: industry productivity, reliability, new investment, and prices. 11.3.2. Price outcomes In the UK, the NETA market model brought a claimed 15% price reduction (on top of the 30% real reduction in 1990–2000)2 . Australia saw prices for larger customers fall 28% in 1996–1999 according to a number of surveys. Prices for smaller customers were reduced by regulators. Although real prices in Australia have edged up recently, they remain considerably below the 1994 pre-reform levels. The easiest and least ambiguous measure is wholesale prices. Compared with a notional $40/MWh (about $50 in today’s money) that was the transfer price between the afﬁliated branches of the state-owned business prior to reform, spot and contract prices have been as shown in Table 11.1 and Fig. 11.5. The Australian Energy Regulator (AER) analysis of ﬂat (baseload) contracts shows no general upward movement.

EBL

Table 11.1. Average prices $/MWh Year

NSW

QLD

SA

1998–1999 1999–2000 2000–2001 2001–2002 2002–2003 2003–2004 2004–2005 2005–2006 2006–2007

3313 2827 3769 3476 3291 3237 3933 3724 3483

5165 4411 4133 3534 3779 2818 2896 2812 2442

15602 5927 5639 3161 3011 3486 3607 3776 3924

SNOWY 3234 2796 3706 3159 2983 3080 3405 3109 3525

TAS

VIC

19038 5676 3992

3633 2635 4457 3097 2756 2538 2762 3247 3629

Source: NEMMCO http://www.nemmco.com.au/data/avg_price/averageprice_main.shtm

2 See for example, National Audit Ofﬁce (2003). The New Electricity Trading Arrangements in England and Wales, http://www.nao.org.uk/publications/nao_reports/02-03/0203624.pdf

Resource Adequacy and Efﬁcient Infrastructure Investment

397

S/MWh 120 80 40 0

ar

9

9

-9

M

-9

p Se

Queensland

0

-0

M

ar

0

-0

p Se

1

-0

M

ar

1

-0

p Se

2

-0

ar

M

New South Wales

03 r-04 -04 r-05 -05 r-06 02 03 p p p- ar- epa a a e M M M M S Se Se S Victoria

South Australia

Tasmania

$/MWh

80 60 40 20 0

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 06 06 07 07 07 07 08 08 08 08 09 09 09 09 10 10 New South Wales Queensland Victoria South Australia

Fig. 11.5. Regional quarterly spot and future prices. Source: AER (Australian Energy Regulator), December 2006. http://www.aer. gov.au/content/index.phtml/tag/MarketSnapshotLongTerm Analysis.

EBL

11.3.3. Changes in Australian operations’ efﬁciency levels Underpinning these real price falls have been large increases in efﬁciency. For example, in Victoria the generators since being moved into a more competitive setting (following corporatization in 1994 and their subsequent privatization) have seen their workforces shrink from about 11 000 to the equivalent of less than 2500. South Australia, the other state that has fully privatized, saw similar improvements in generators’ labor productivity. Government-owned generators also improved and even Queensland (partly private-owned), which had long been better managed than other states’ industries, saw a doubling in productivity. Figure 11.6 illustrates the different state outcomes. Improvements were also seen in the level of reliability of the power stations, especially in Victoria and NSW, the two state systems that were previously performing poorly (see Fig. 11.7). Improved productivity was registered in other areas of the industry, including the regulated distribution businesses. Again, this was most marked in the privatized systems in Victoria than in the government-owned systems. It seems likely that part of the reason for this is the closer commercial focus of private businesses. There is also some residue of political appointments to the corporatized businesses’ boards. Ten years ago, the CEO of the largest of the NSW distribution businesses attempted a reorganization to capture the same labor-saving gains as his counterparts in the Victorian privatized businesses. Its government-appointed board of directors responded by sacking him. This has become less frequent, though in November 2006, the NSW State Treasurer sacked the long-serving

Competitive Electricity Markets

398 45 40 35

1990/91 1996/97 2000/01 2003/04

30 25 20 15 10 5 0 New South Wales

Victoria

Queensland South Australia Western Australia (Western power)

Fig. 11.6. Generator Labour Productivity (GWh/employee). Source: ESAA, Electricity Gas Australia 2006.

95 1990/91

90

1996/97

EBL

85 80

1999/2000 2004/05

75 70 New South Wales

Victoria

Queensland

South Australia

Tasmania

Western Australia (Western power)

Fig. 11.7. Power stations’ availability to run. Source: ESAA, Electricity Gas Australia 2006.

Chairman of the state’s transmission business allegedly because he would not agree to an appointment of a politically favored director3 . Figure 11.8 illustrates the trends in terms of customers per employee. 11.3.4. New investment outcomes Even though the competitive environment has meant low prices, windows have opened where ﬁrms have spotted (or thought they spotted) opportunities to expand. Though the presence of government-owned facilities may well be distorting new provision – a point that is addressed later – the market has, to date, not only produced lower 3

Salusinszki, I. (2006).Sacked for rejecting union mate. Australian, 16 November.

Resource Adequacy and Efﬁcient Infrastructure Investment

1200

399

1994/95 1998/99 2003/04

1000 800 600 400 200 0 New South Wales

Victoria

Queensland

South Australia

Tasmania

Western Australia (Western power)

Fig. 11.8. Distribution businesses: Customers per employee. Source: ESAA Electricity Gas Australia 2006.

prices but also resulted in capacity increases in line with demand. Table 11.2 shows new capacity. In terms of average costs of new electricity increments on the eastern seaboard, coal is $35–$40/MWh and gas about $45 though this is based on a gas price that is at present less than half of that in the US. Capital costs are illustrated in Table 11.3. For CCGT plant, the cost $1000/kW to $1250/kW in 2006 represents a considerable increase from that which was estimated at under $900/kW in 2005. The increase is attributed to temporary cost increases in steel and other materials as a result of a surge in demand in China and India. For coal-based generation, a cost of about $1250/kW is indicated for Kogan Creek if there were no mine development costs.

EBL

Table 11.2. New capacity 2000–2006

Redbank Bairnsdale ValleyPower Somerton Laverton Loy Yang Oakey Millmerran Swanbank E Tarong N Kogan Creek Braemar Hallett Pelican Point Ladbroke Quarantine

State

Capacity (MW)

Type

Ownership

NSW Vic Vic Vic Vic Vic Qld Qld Qld Qld Qld Qld SA SA SA SA

150 92 300 160 312 236 282 852 360 450 750 450 220 500 80 100

Coal Gas Gas Gas Gas Coal Gas Coal Gas Coal Coal Gas Gas Gas Gas Gas

Private Private Private Private Govt. Private Private Private Govt. Govt./private Govt. Private Private Private Private Private

Source: ESAA, Electricity Gas Australia 2006.

Table 11.3. Capital costs of new plant Power station

Proponent

State

Cost in A$ million

MW

Cost in A$/kW

Type of plant

Fuel

Comments

Cockburn Braemar

Western Power ERM and Babcock Brown CS Energy

WA Qld

250 340

240 450

1042 756

CCGT OCGT

Gas Gas

Completed Nov 2003 Under construction

Qld

1200

Coal

WA

400

Laverton North Kemerton

ERM and Babcock Brown Snowy hydro Transﬁeld

ST (dry cooled) CCGT/ST

Vic WA

Gas Gas

Wagerup

Alinta

WA

Under-construction (Includes coal mine development) Under construction (Includes 160MW ST) Under construction Under construction (Includes gas connection) Committed

Kogan Creek Kwinana Newgen

Source: ACiL Tasman for NEMMCO, October 2006.

EBL 750

1600

320

1250

150 250

320 260

469 962

OCGT OCGT

245

324

756

OCGT

Gas

Resource Adequacy and Efﬁcient Infrastructure Investment

401

11.3.5. The drought of 2007 Australia, a ﬂat and dry place, has never had much hydro in its generation sector (less than 10% of energy production) and therefore its power prices were never considered sensitive to rainfall. The ﬁrst months of 2007, however, proved that perception wrong. A severe drought affected the entire eastern seaboard and curtailed not only hydro generation, but also a number of inland coal-ﬁred plants requiring cooling water. The impact of this disturbance on spot and forward prices during early 2007 was quite dramatic, with both approximately doubling in a space of 3 months. This is easily explained in the energy-only market, as the withdrawal of “free” hydro and cheap coal energy must be replaced, even at offpeak times, with a more expensive gas turbine plant. Understandably, there is some political repercussion, principally from advocates of those large customers who have to purchase new supply contracts. Fortunately, most customers, including small ones have contracted long-term well before the current situation, and, in turn, their retailers have contracted with generators. Thus, the political pressure is not overwhelming upon governments and regulators, who, at least as of December 2007, were not undermining the market. At the same time, the water issue is very sharply biting into the proﬁts of those generators affected by it. They are demonstrably taking on innovative responses, such as purchasing high-priced water, and investing in previously uneconomic conservation. All this is occurring without any regulatory intervention or “guiding hand,” beyond the clear proﬁt motive of such a high opportunity cost. Figure 11.9 illustrates the price surge that had taken place. Forward prices from 2008 are starting to return to more normal levels, though they remain somewhat higher. This may be due to concerns that the current drought is part

EBL

NSW

QLD

VIC

90 85 80 75 $/MWh

70 65 60 55 50 45 40 35 30 Jul 06 Aug 06 Sep 06 Oct 06 Nov 06 Dec 06 Jan 07 Feb 07 Mar 07 Apr 07 May 07 Fig. 11.9. Flat forward price curve – All regions. Source: AGL.

Competitive Electricity Markets

402

of a “global warming” pattern and uncertainty regarding future carbon taxes as well as some tightening in supply. 11.4. Reliability and Capacity Reward 11.4.1. Reliability – Actual performance As seen in Table 11.2, considerable new capacity has come about since the market was conceived, roughly in parallel with the national growth in demand4 . Previous over-supplies in New South Wales and Victoria have eroded, whilst previous under-supplies elsewhere have been remedied and, in Queensland’s case, reversed. Actual performance has been excellent in terms of “reliability,” as the NEM deﬁnes it, meaning the overall adequacy of generation supply. (Load losses due to other causes, such as local distribution network interruptions or transmission stability problems are not avoidable through generation investment and therefore excluded.) The ﬁrst eight years of the NEM have seen the following percentages of demand interrupted due to lack of reliability: • • • •

New South Wales, 0.0001% Queensland, 0% South Australia, 0.0025% Victoria, 0.0101%5 .

In all cases, except Victoria, this would be considered a very successful reliability outcome for a ﬁrst-world power system, and immaterial compared to typical local distribution outage losses of around 0.02%. The Victorian amount in turn is derived entirely from a power station strike in 20006 .

EBL

11.4.2. The reliability standard A forward-looking reliability standard is used for the NEM based on a minimum level of projected reserves for the year ahead as deﬁned by the Regulator7 . 4

Peak demand has grown by 4510 MW and supply by 5138 MW from 2000–6. See Australian Energy Market Commission (2006). Annual Electricity Market Performance Review: Reliability And Security. Available at: http://www.aemc.gov.au/pdfs/reviews/Annual%20Electricity%20Market %20Performance%20Review%20-%20Reliability%20and%20Security %202006%20Report/aemcdocs/ 001Draft%20Report.pdf 5 Australian Energy Market Commission (2006)Annual Electricity Market Performance Review: Reliability And Security. Available at: http://www.aemc.gov.au/pdfs/reviews/Annual%20Electricity %20Market%20Performance%20Review%20-%20Reliability %20and%20Security%202006%20Report/ aemcdocs/001Draft%20Report.pdf 6 Interestingly, while strikes in the power industry commonly afﬂicted Victoria pre-market, the strike in the early days of the market was caused by a legacy of that culture and has not been since repeated –arguably, an outcome of the clear ﬁnancial incentives upon generator performance created by the market. 7 The NEM’s reliability standard is set by the “Reliability Panel,” an independent body with broad membership and expertise. But it also accidentally beneﬁted from ﬁrst being implemented at a time of turmoil, where the government businesses were being divided up and vested interests were unclear, allowing something of a free reign to economic technocrats. This may explain why it is at a more moderate level than that demanded in some other markets.

Resource Adequacy and Efﬁcient Infrastructure Investment

403

Where this standard is deemed to be breached in the short to medium term, NEMMCO is obliged to intervene in the market. As with other such interventions, this carries the potential, if used unwisely, to undermine the market reward function – a matter discussed in Section 11.4.4. To date, this power in Australia has been used sparingly, largely because the minimum reserve standard is softer than that in most other jurisdictions. The NEM’s standard is an output-based standard: a measure of customer energy actually at risk. This is set at “an expectation of no more than 0.002% of energy unserved over time.” It means that no more than 1 in every 50 000 light bulbs should go out due to generation shortfall. Or, to put it another way, customers will suffer, on average, no more than 10.5 minutes per year of interruption. This compares with an average 100 minutes per year interruption in Australian suburbs due to local distribution faults. The requirement was created based on customer surveys which suggested the typical customer values reliability in the order of $20 000 to $30 000/MWh8 . Considering the cost of providing peak generation capacity to meet the extreme peak of the demand shape, the cost of supplying the last 0.002% is actually greater than its customer value. Using power system simulation to convert this unserved energy target to a deterministic reserve margin for a one in two years peak is equivalent to around 10–15% reserve margin. The resultant standard compares with standards as high as 25% in many jurisdictions around the world, standards which, if worked backwards through the simulations, would mean that customers energy targets unserved are as low as 0.0002%, or about 1 minute in a year, and they are valuing reliability at $100 000/MWh! Kema Consulting (2005) notes that the Australian approach is at the low end of international standards. Were those levels of forecast reliability demanded of the NEM, a much higher price cap would have been required. Indeed, it may be that any energy-only market will struggle to deliver those reserve margins, not because the energy-only market has failed a prerequisite, but that it is in fact simply correctly demonstrating that customers do not value reliability so highly.

EBL

11.4.3. The price cap and generator bidding freedom Discussion elsewhere in this chapter emphasizes the need for a genuinely high price cap for an energy-only market like Australia. The notion of a price cap reﬂects the view that electricity has major differences from other markets in view of the instantaneous nature of the commodity, and the physical inability to link customers immediately to its price. Setting a price for such interventions is designed to allow a very large pain to those illprepared (and conversely a great opportunity for those who can help), but one that does not immediately result in a systemic ﬁnancial collapse of market participants. The current $10 000/MWh that has been in place since 2002 seems to be facilitating an adequate level of investment and relatively low customer prices. The price cap is reinforced by another mechanism, the “Cumulative Price Cap” which is set at $446/MWh for an average of prices over a rolling week. Price is then capped at $50/MWh and $100/MWh peak. The rolling price cap has not been reached in the period since 1998 when the market commenced, though it has got close on a couple of occasions in situations which were not actually threatening the market’s ﬁnancial collapse, indicating that it is too low. 8

See documents such as Victorian Energy Networks Corporation (2002). Value of Customer Reliability Report.

404

Competitive Electricity Markets

Whilst a generator market cap is defensible as a measure to prevent market ﬁnancial collapse, if the objective is instead driven by consumer price protection, then the energyonly market is probably doomed as consumers or their agents will prefer to ride upon this much cheaper protection than invest in supply. A market cap is always a departure from a pure market approach, and the lower is the cap, the more vulnerable the market becomes. In Australia’s case the $10 000/MWh cap is considerably below most estimates of the Value of Lost Load (over $30 000/MWh.). Even so, it appears that prudent retailers seek to insure themselves to the very peaks of their forecast demand and generators invest well before any shortfall manifests itself. Retailers’ apparently irrational prudence is driven by the fact that generators have the freedom to exercise their market power. Indeed, it is not uncommon for large portfolio generators to shadow the $10 000/MWh price cap for as much as 20% of their capacity and, therefore, high prices may occur well before actual interruption. What is even more impressive is that there is no legal or political sanction for this behavior, so it constitutes a genuine threat to those who expose themselves. During 2001, a “good faith” rule was inserted into the market. This is purely a mechanism designed to prohibit intentional deception. In theory, in an energy-only market, a generator can confuse its competitors by changing its bids at the last minute. This rule prohibited last-minute changes where its own or market conditions were unchanged, but in no way does it attempt to limit their market power. Indeed, acceptable public reasons for lastminute bid changes include “change in market price/volume trade off.” After three years of the rule being in place there have been no prosecutions under it. 11.4.4. Capacity reward and intervention

EBL

Uncertainty about the adequacy of market remuneration has led to questions about the appropriate incentives to invest in new generation. These questions spawned several answers. With the NEM as a pure energy-only market, reliability sufﬁcient to satisfy the many stakeholders is a likely outcome. Of course, Australian governments are no less fearful than others of the unknown in relation to blackouts, and this adds a complicating set of regulatory factors. In the NEM, the regulatory responses are centered on the concept of Reserve Trader. This overrides the market supply when the market operator decides that there is insufﬁcient supply forthcoming from the market in the foreseeable future. The problem, other than that of explaining how a public sector body is more likely than the market to predict supply and demand conditions, is that the Reserve Trader as a concept has internal inconsistencies. If the public agency (called NEMMCO in Australia’s case – the market/system operator) considers there to be insufﬁcient supply, it must contract for that supply. In doing so, it must either: •

move into the market and contract supply at a higher price than the supply was able to get from real customers; or • build its own capacity. In an attempt to avoid undermining the market, NEMMCO is limited to contracting for reserves no further than about six months ahead. Due to practical difﬁculties, this largely excludes new entrant generators. Thus, it is likely to only get mothballed supply or demand-side opportunities, and it will contract for this by providing a higher price than is available in the general market. While the consequential price increases may not

Resource Adequacy and Efﬁcient Infrastructure Investment

405

be serious, they do raise costs to customers, thereby defeating the purpose of the market model. More than this, the process will encourage ﬁrms or demand-side suppliers to hold back offering contracts to the market in the hope that the government will offer them a better price. While such data is conﬁdential, it is likely that several customers contracted in this way during 2005 had previously participated in the market for a market-based return. This has a snowballing effect in creating even greater apparent shortages and can start a process that will unwind the market itself. An example of such an outcome has been reported by Joskow (2006): “In New England, the amount of generating capacity operating subject to special reliability contracts with the ISO has increased from about 500 Mw in 2002 to over 7000 Mw projected for 2005 (ISO-New England (2005), amounting to over 20% of peak demand.” Such a proportion of the market subject to administration must start to undermine the commercial market as a whole. If the reserve power agency were to hold its own capacity to be used only in special circumstances [e.g. when the price exceeds the spot market cap (VoLL) or an agreed period of time], this is simply an added insurance on VoLL and a drain on the market. Of course, if the reserve capacity were to be used more liberally than this it would undermine investment incentives and contribute to supply shortages in the future. Another answer to capacity shortages is a capacity payment offered in addition to the energy price. Some incumbent Australian generators, dismayed at the very low prices they are seeing, favor this. Against this, it has to be recognized that if additional payments are made for supplying energy for one set of reasons, compensating reductions will occur with related payments as ﬁrms jockey for revenues that cover their costs. Moreover, experience has shown that where supply is ample, the capacity price will be bid down perhaps to negligible proportions. In this respect, Adib, Schubert, and Oren in Chapter 9 of this volume discuss what they call the “bipolar nature” of capacity markets with price being zero where capacity is ample and inﬁnity where it is short9 . Where there is already some market imperfection, as appears to have been the case in the original England and Wales market, the capacity payments may become high as ﬁrms use market power to bid them up. A single price or the addition of a capacity charge as the most appropriate way forward must, however, remain one of the open questions in market design around the world. In California, the issue is being reviewed once again but the California Public Utilities Commission staff (2005) is very much in favor of a capacity charge. They argue that electricity is different because of its near-total demand inelasticity – the inability to selectively supply people – and thus have a differentiated reliability; and they note that a price cap, which they see as inevitable, also means less than ideal conditions for individual risk management. This does not seem to be borne out by the experiences of the energy-only market that is in place in Australia. An energy-only approach appears to be superior to all the reﬁnements that have been tried elsewhere. It places the onus on commercial parties to cover their future positions in the knowledge of their customer bases and future demand shifts. Suppliers and retailers develop their own reserve trader through contracting in ways that give them adequate insurance for mistakes and uncertainty. It has served the Australian market well in terms of incentives. New capacity in generation has kept pace with requirements.

EBL

9 Adib, Schubert, and Oren also develop a procedure for a capacity payment mechanism where the energy-only market might not operate (because of price caps and other regulatory interventions). The authors see their proposals as a transition to a more comprehensive energy market.

406

Competitive Electricity Markets

Among the measures that have ensured the energy-only market operates successfully in Australia are a relatively high reserve price of $10 000/MWh, which places considerable pressure on retailers to forecast and balance demand accurately and to contract for their customers’ future needs. This, in turn, provides incentives for generators to deliver the necessary capacity. Further assisting this is a relatively open and active retail market and relatively unimportant customer price caps (soon to disappear entirely). All major customers have been free to seek their own retail supplier for many years, and the household and small business consumers are likewise mainly freed from dependence on their original retailer.

11.4.5. The role of the electricity retailer 11.4.5.1. Retail competition A competitive retailing system is a bridge between the producer and the consumer and provides, by seeking out customer needs and arranging for their supply, important signals for new production and speciﬁc sorts of new production (electricity that is peak, offpeak, green, etc.). One (imperfect) measure of retail competition is the degree of customer churn. According to Grey et al. (2005), in 2004, Great Britain, Victoria, and South Australia were the “hot” markets for retail switching, with only Texas in the US ranked in their next category, “active.” Littlechild10 estimated the numbers of residential customers with non-incumbent suppliers as ranging from 43% in the UK and 33% in Victoria to very low shares in US states other than Texas (where the share was 24%). August 2006, data for Australia indicates that 67% of Victorian and 27% of NSW customers had switched from their host retailers11 . Customers with a contract other than with their host retailer comprised 42% in Victoria and 18% in NSW. In South Australia, 64% of customers had shifted out of the default contract (there was only one original retailer) by the same date. The lower level of “churn” in NSW is due to two factors. First, there is a mandatory insurance scheme for small loads. Though this is being discontinued it places out of state retailers in a less favorable position to hedge against risk since the intra-state retailers have a lower de facto peak price. In addition, the level of retail price cap bites earlier than in Victoria and South Australia, meaning that a larger proportion of residential buyers are, in effect, unable to obtain commercially a better deal than the government has mandated their retailer offers them. Another feature of the NSW arrangements is that consumers are able to return to a standard tariff, should they wish. Such fall-back tariffs offering a one way bet have unwound retail deregulation in a number of US jurisdictions. Australia has seen the emergence of a number of new retailers, some of them very small, and in several cases their entry has been successful. There are concerns that full retail competition can bring instability where retailers have taken unreasonable risk and then left the market leaving other retailers to continue supplying their customers. This apparently occurred in Texas in 200312 .

EBL

10

Littlechild, S. (2005). Beyond regulation, Beesley Lectures on Regulation, Institute of Economic Affairs, London, http://www.iea.org.uk/ﬁles/upld-article94pdf?.pdf 11 This may contain an element of double counting since it includes customers who have switched more than once; on the other hand, it excludes customers who have moved off the default tariff but remained with their existing retailers. 12 Personal communication Shmuel Oren, January 2007.

Resource Adequacy and Efﬁcient Infrastructure Investment

407

The requirement of a retailer of last resort is certainly an area where public policymaking to protect the small consumer has the potential to unwind the proper forces of an energyonly market. The Australian pool mechanism demands of all retailers quite onerous credit assurance for both pool and networks. This has the potential to be inefﬁcient, but can be overcome with voluntary settlement offsets with generators. The credit requirements place pressure on retailers to ensure that they are prepared for price volatility. This should mean that retailer bankruptcy is very unlikely, or that, rather than short-payment, it will be inability to get assurance that will lead to a retailer’s forced exit – which means it can be managed more effectively and the customer accounts are likely to be sold to a willing, and more prudent, buyer. In the event that a retailer becomes bankrupt, Australia has a retailer of last resort for smaller customers who comprise half the market. The liability is on foundation “host” retailers to absorb these customers. The actual arrangements vary between the states. In Victoria, the government allows the host to immediately replace their tariffs with a much higher price than the typical competitive level (10–20%). In some respects a last resort retailer fulﬁlling contracts of a failed competitor is not different from many other industries. For example, airlines will usually take up emergency cases of stranded passengers when a carrier goes bankrupt and ceases to operate. 11.4.5.2. Retailers as drivers of efﬁciency As retail margins are only about 5% of cost, some are perplexed by the prominent role given to the retailer in the judgments about the liberalization of markets and, implicitly, about how they correspond to consumer beneﬁt. A major push at one time was to have tariffs set on a “pass-through” basis. However, retailers focused on customers and suppliers in a competitive situation ensure a sound alignment between the two. Competition is, fundamentally, a discovery process, whereby the competitors set out to ascertain the needs of customers, where those needs are not well deﬁned by the customers themselves. Evidence of such poor alignment in the centralized system can be seen with the excessive priority on base-load seen throughout Australia, which led to a major surge in new peaking capacity once competition was in place. This has meant a bonus of much better reliability at lower cost. Competitive markets provide particularly strong incentives on retailers to search out the lowest-cost supplies and match these with customer demands. This is particularly so in Australia’s case where the wholesale cost of electricity can rise to $10/kW hour compared to its normal price of about 4 cents per kW hour. With a “pass-through”-regulated tariff retailers would gain no beneﬁt in seeking innovative and highly competitive supply contracts. The economic beneﬁt of such innovation would simply pass to the customer whilst the implicit costs, such as greater risk, would fall on the retailer. The only incentives such a retailer faces to attain efﬁcient supply would be artiﬁcial ones set by the regulator of the pass-through process. These would always be out of date, out of touch with the customer and conservative. One outcome would be that retailers supplying customers who may have a more peaky demand proﬁle would not have an incentive to ﬁnd suppliers who are the most efﬁcient supplier of such a load shape. They would also be indifferent in seeking out such customers rather than others with ﬂatter load proﬁles. The associated suppression of cost-reﬂective price variations is likely to rebound on the efﬁciency of the entire investment chain, including the highly capital-intensive sectors. Retail competition also offers other beneﬁts. For example, it facilitates a variety of different product offerings. Among these have been “green” power packages, obtainable

EBL

408

Competitive Electricity Markets

by those prepared to pay a premium for this form of electricity supply. It also reveals the extent of the voluntary demand for such products. Competitive retailing has also meant experimentation with new marketing tools. Victorian retailers have successfully experimented with direct debiting of customers’ accounts and there have been experiments in combining energy with other retail activities. There are clear dangers in overriding the forces of competition, dangers that intensify with the length of time the controls remain. For retailers themselves, setting prices too low will require cross-subsidies. Aside form their general inefﬁciencies, these will bring about an unraveling of the market balance because it will prove increasingly difﬁcult for the regulators to set ﬂexible prices which are cost-reﬂective and do not leave the retailer in a revenue deﬁcit. Financial distress among retailers ensuing from such price caps is likely to be an early manifestation of an impending crisis perhaps culminating in California-style collapse. This aside, the price suppression involved in regulation distorts the signals for augmentation in new generation. At best, this will bring inefﬁcient balances between peak and offpeak and at worst, it will lead to supply inadequacies. Many are keen to see “smart” meters being installed to allow time-of-day measuring of power use by small consumers who account for half of the load. This would allow pricing for those using air-conditioners during peak hours to match the higher supply costs involved. It would drive peak load reductions and correspondingly lower charges to other customers. The overall beneﬁt of this turns on the potential cost saving against the installation costs of the meters themselves. These sorts of metering have not been very successful in facilitating load shaving in the large business markets which have long had the metrology and controls to facilitate this. Experience suggests that regulatory interventions to force the pace of change should be subjected to critical assessment.

EBL

11.5. Some Key Issues and Fragilities 11.5.1. Global warming and new generation Australia has perhaps the cheapest primary energy in the world available in major quantities. Coal from Queensland and parts of NSW is abundantly available for conversion into electricity at $40/MWh virtually forever. Brown coal in Victoria is available at a similar price. These prices are less than a half of those in Japan and considerably below those of the EU and most of the US. Coal at $40 is half the price of wind energy (the costs of which are ﬂattered by its inherent unreliability) and the cheapest nuclear option is about 30% dearer including the (relatively low) disposal costs. Figure 11.10 illustrates costs. A greenhouse tax would be a great equalizer. Figure 11.11 illustrates the costs with a carbon tax or tradeable right set at the Stern Report’s (2006) $A130 per tonne of CO2 . With such an imposition, natural gas becomes a bit cheaper than coal, though this might be offset by a rise in its price, which in Australia is less than half that of the US. The big movers (or stayers) are nuclear and wind. Wind on the assumptions given becomes cheaper than coal in Victoria and NSW, though its role can never be to supply more than about 10% of the load at almost any conceivable price and with the most heroic assumptions on future improvements. Nuclear though assumes the leading position. Uranium is relatively abundant and comprises only a small share of costs, the bulk of costs coming for plant. Doubtless, these costs are also inﬂated by over-engineering to cater for hysteria over safety matters. This, and the fact that relatively few new nuclear plants have been built in recent years, means that the price might even be reduced below the levels indicated by current studies.

Resource Adequacy and Efﬁcient Infrastructure Investment

409

Cost per MWh $80 $70 $60 $50 $40 $30 $20 $10 $Black coal Black coal Brown coal NSW Qld

Natural gas

Wind

Nuclear

Fig. 11.10. Electricity costs. Source: Authors based on several sources.

$200 Present costs

$180 $160

Costs with greenhouse tax

$140 $120 $100

EBL

$80 $60 $40 $20 $Black coal NSW

Black coal Qld

Brown coal Natural gas

Wind

Nuclear

Fig. 11.11. Costs of different electricity sources.

So in a carbon-constrained world, there is a means of abundant and reliable electricity supply that will allow existing consumption at only a modest increase in costs. However, even this is insufﬁcient to provide the emission savings of 60% or so sought by the Stern Report. Moreover, Australia has no advantage in nuclear. Australia’s advantage is in cheap fossil fuel-based energy. Abandoning that advantage, even progressively, will not only mean far higher capital costs but also elimination of the nation’s comparative advantage in energy-intensive industries. It will, therefore, at a minimum, entail a considerable industrial restructuring. Moving to the prospect of an energy tax or a tradable right to emit introduces a considerable uncertainty in new plant development. It is notable that the new large-scale base-load coal plants in recent years have been built by government entities; hence, the government is taking the regulatory risk of some ex post facto new imposition. Greenhouse mitigating activity does infer some additional commercial risk, which may add yet a further uncertainty about future supplies.

410

Competitive Electricity Markets

11.5.2. Government ownership Over half of Australian electricity generation capacity is in government hands. Although all government generation businesses operate under normal company law with directors that are independent, the fact that the directors are appointed by the state governments gives rise for some concerns regarding their independence from political processes. In NSW, there are suggestions that the government generators are restrained from major new investments by a government conscious of previous excesses in development. It has been suggested that the opposite problem prevails in Queensland, the state with the fastest-expanding load. In the ten years from 1990, the state built only one major power station and its precarious balance of supply and demand was immediately revealed once the electricity market went live during 1998. Shortages that had been hidden were immediately reﬂected in wholesale prices that were double those of the southern states. Remedying this was essential. And since Queensland, along with parts of NSW, probably has the lowest-cost abundant quantities of coal in the world, remedying this was straightforward. Over the six years to the end of 2006, Queensland increased its electricity generating capacity by a quarter, adding 3300 MW, 60% of the new capacity within the Australian National Electricity Market. As part of its initial catch-up in capacity, the government encouraged private investment to enter the market. A Shell-dominated consortium built the 850-MW Millmerran power station in 2002. That consortium also took a half share with the government’s CS Energy in the 920-MW Callide C station that was completed a year earlier. The 450-MW Tarong North, started in 2000 and completed in 2003, also had a mixture of government and private funding. A further government-owned major major station was announced in May 2004. Soon after entering the market, Shell clearly felt its investment had turned sour and steadily sold down its interest. The ﬁnal one-quarter share went for $US226 million in December 2003 to China Huaneng Group. Perhaps the Chinese bought well, but the transaction valued investments that had cost some $2.2 billion at only $1.2 billion. In this respect, the danger is that investments undertaken on non-commercial terms using government funds will undermine all investments. Private sector businesses argue that the Queensland government, having enticed investment into base-load power, has then accepted non-commercial rates of return from the power stations it owns. Some credibility to this claim has been given by statements by the Queensland Energy Minister that he sleeps easier if he has 25% surplus generation capacity. However, a corollary of such a supply margin is a collapse in prices and in the value of assets. Although there are relatively robust transmission links between Queensland and New South Wales, the NEM state to the south, spot prices in Queensland have been 30% lower. And prices in NSW were themselves considerably reduced by the export of surplus power from Queensland. Even so, the higher prices in NSW have encouraged the Queensland government to seek an augmentation of transmission capacity to take advantage of those prices. But stateowned NSW generators see their market as already being infected by surplus Queensland capacity and low prices from oversupply. If there is a loss of proﬁts by private investors caused by government accepting sub-commercial rates of investment return, this risks creating a vicious circle under which all future investments will be state funded.

EBL

11.5.3. Ensuring adequate transmission capacity Among the most contentious issues have been and remain the ability to supply the right amount of transmission capacity. Many have argued that transmission should be provided

Resource Adequacy and Efﬁcient Infrastructure Investment

411

without the need for this to be fully commercial in the sense normally required of interventions throughout the economy. In this respect, positions have changed little over the past decade. Thus Hogan (1998), although injecting a market-type mechanism into transmission provision, saw that “Grid expansion and pricing would continue to present a need for regulatory oversight, but the existence of workable transmission congestion contracts would substantially simplify transmission investment decisions.” (p. 28.) In Australia, London Economics (1999) argued that short-run congestion could recover at most 25% of transmission costs and, by inference, transmission must be supplied on a regulated basis with mandatory charges. L.E. estimated recovery in US markets was 5–20% of costs, with the highest level of recovery they could identify being Queensland at 24%. Similarly, the US National Energy Policy document of May 2001 argued that, “The transmission system is the highway system for interstate commerce in electricity. Transmission allows the sale of electricity between regions. In a particular region, transmission can be a substitute for generation, allowing that region to import power that otherwise would have to be generated within that region.” But while it recognized the importance of incentives to augment the transmission system, it saw these as being rate- based with a regulatory backbone and did not contemplate the implications of this for its substitute, generation, and the consequent market distortions. Attempts to place transmission provision on the same basis as generation has proven difﬁcult. Australia’s experiments with merchant transmission have not been successful. The new entrant Transenergie13 has now opted for regulated status of its lines and sold out of Australia. This may reﬂect the intrinsic inability of such facilities to earn sufﬁcient return because of lumpiness and externality issues. Others would argue that such matters are equally prevalent in power stations: they are normally too large for their immediate requirements. With regard to externalities, it is argued that these are too great to allow proﬁtable merchant transmission, because the beneﬁts of lower prices (actually arbitraged prices) accrue to all and not only to those paying for the asset. This has spurred proposals to reward new transmission investors from gains made by consumers (see for example, Haydon and Michaels, 2006). However, the effect of transmission augmentation is not markedly different from the situation concerning a new generation facility that will tend to suppress the price of all delivered electricity in its interconnected region. Few would argue that by analogy, all generation should, therefore, be government-owned or subsidized, even though many argue for a form of general overhead support in the form of capacity payments. The fact is that supply across the economy is seldom unaccompanied by some externalities. If transmission is provided free or at regulated prices, this may discourage a more rational and lower cost development of new generation. The tradeoff between nearby and remote generation (via transmission) is uniquely critical for Australia, where distances between load centers and therefore the cost of transmission are very large, and fossil fuel sources are relatively inexpensive and quite widespread. The danger is that links which are ﬁnanced by a compulsory charge on the customer might lead to incentives to site generation in places that are distant from major markets. If someone else is paying for transmission, the rational generation business will be indifferent to its costs, thus distorting the efﬁcient tradeoff transmission costs and generation costs.

EBL

13

A subsidiary of Hydro Quebec.

412

Competitive Electricity Markets

Associated with the claim, that transmission would be inadequately provided in the absence of it being made subject to regulated support, is the contention that a transmission line has market power and its prices should be regulated. However, for the most part, transmission inter-ties or interconnects offer no more market power than that of a signiﬁcant generator portfolio. Inter-ties in Australia can account for some 35% of supply (Victoria to South Australia) but normally provide much less than this. Their market power is conﬁned to inﬂuence over those wishing to export, and such ﬁrms are normally capable of writing contracts to cover any vulnerabilities they foresee. How best to allow expansion of transmission, especially in terms of the regional linkages, has been the subject of a heated debate in Australia. An uneasy compromise is presently in place for transmission under which regulated links will be permitted as long as a net market beneﬁt is judged by the regulator to be the outcome and as long as the proposed link is the best of a range of feasible alternatives. This, however, remains dissimilar from the decision-making structure that is seen in the generation sector or in markets more generally, since it may incorporate some to the network beneﬁt externalities which a comparable investment in a new generator would not capture. The competing solutions that generation and transmission often offer mean disputes about the merits of a new transmission solution are likely to remain. These are illustrated by pressures from the Queensland government to augment transmission links following the state’s capacity increases driving down prices below those in NSW. This might be regarded by others as facilitating dumping by having expanded capacity ﬁnanced as a regulated link because most of the costs fall directly on consumers. As Michaels (2006) argues, establishing a market in a condition of supply surplus is a relatively straightforward matter. Ensuring its ongoing development requires an appropriate structure. Michaels regards the separation of transmission and generation as potentially fatal. He says:

EBL

Studies in the 1980s and 1990s almost invariably concluded that vertical integration produced efﬁciencies that would be lost in a breakup. These economies of integration applied to both the generation-transmission interface and to the ownership of generators and fuel supplies. This scholarship was almost totally forgotten as California and other states began to restructure their power industries in the mid-1990s. The matter of establishing an appropriate regime for transmission development is again being considered before the Australian Energy Market Commission (AEMC). The AEMC recognizes that investments may be inappropriately located because of the charging approach. It favors prices being set on the basis of short-run marginal costs, which it argues is supported by economic theory and competitive market experience. This is subject to a great many caveats. Importantly, prices set on the basis of marginal costs are not found in many markets – they are characteristic of markets under stress (for example, where there are few suppliers engaged in a “price war”) or facing long-term decline (so that sunk costs need not be recouped). Even the market for highly perishable goods like vegetables seldom sees produce offered at marginal cost and only then is this seen at the end of the trading day. The AEMC recognizes that if charges are set to meet short-run marginal costs and there is spare capacity, consumers may locate too far away from generation, especially if reliability standards are in place to fortify the initial decision. It considers that prices based on long-run marginal costs may lead to inefﬁcient bypass. This leads it to support the notion of efﬁcient discounts being offered which may be recouped by de facto surcharges on other customers. It is likely that the conditions

Resource Adequacy and Efﬁcient Infrastructure Investment

413

under which these would be permitted would be accompanied by protracted and heated negotiations. The practise in Australia is to charge the customers for the transmission use, rather than generators. Generators, however, do not have a property right to the transmission to the major hub. This means that a new generator with costs and a consequent bidding strategy lower than that of an incumbent generator would force the latter off the line once it was at full capacity. This might mean an alternative supplier with a higher total cost (including transmission costs to the major node) would replace the incumbent generator. This is demonstrated in Fig. 11.12. If generator B locates next to generator A, the latter is constrained out and replaced by the higher-cost generator C. Costs are $1000 higher. Generators A or B may have incentives to build additional transmission capacity but only if they can be assured of some exclusivity or some priority in its use. A customer coalition would also be willing to ﬁnance such an investment but the transmission business may face no such incentive, while generators B, C, and D would prefer the augmentation not take place since they are beneﬁciaries of the higher price set by generator C. Allocation of a form of property right would bring about the optimal investment without the rancor of a series of bureaucratic hearings and extensive lobbying. A new generation unit or an expansion of an existing unit should be required to pay for any augmentation needed to allow its power to be transmitted. This, implicit in which is some form of nodal pricing, gives a better market signal than if the determination is left to a regulator or to a transmission business since it allows the transposition of commercial forces for those that are actually or mainly controlled. As AGL (2005) argued, “Applying deep connection charging to generators at the time of connection would allow the network costs to be included in their decision process on location and allow for appropriate development of networks to efﬁciently transfer power from generators to customers.” With rights over their current levels of service, existing generators have options about augmentation that ensure the full costs of their decisions are taken into account. They may also downsize by selling part of their carriage rights to a new player, thereby avoiding wasteful duplication of capital.

EBL

GenC $50/MWh Cap 500 MW GenD $5/MWh Cap 1000 MW

GenA $30/MWH Cap 1000 MW Line capacity 2000 MW

Line capacity 1000 MW Load 2500 MW GenB $10/MWH Cap 1000 MW Fig. 11.12. Illustration of four generators and two transmission lines serving a load.

414

Competitive Electricity Markets

This is a means of applying a market solution based on a form of property right to the creation of new transmission capacity. It can, if it proves to be practical, resolve the provision of new capacity by taking it away from the artiﬁcial markets that regulators construct in permitting new transmission and levying consumers accordingly. In that way a similar driver is put in place to that of new generation provision and, in principle, a more consistent set of investment incentives are established, thereby avoiding waste and gaming of the regulatory system. 11.6. Conclusions Australia, like the UK, has evolved a market with very little regulation and, importantly, wholesale price caps that are either absent (in the case of the UK) or relatively high (in the case of Australia). Even with market interventions, the outcome has been very satisfactory in terms of serving consumer demand and ensuring resource adequacy. In Australia’s case, those interventions include consumer price caps (diminishing in importance and soon to disappear); a Reserve Trader (sparingly used and never having made any difference in the event); mandatory generator/retail contract hedges in NSW (a semi-protectionist device which is soon to be abolished), and the commercial risks of government owning almost half of generation (though corporatization of their boards means they no longer are pure political instruments). Many are concerned about market abuse. The more independent suppliers there are, the better this is avoided, but even so abuse is not important – indeed, it is necessary in thin markets like Australia’s. Almost all generators in Australia bid some part of their capacity at very high prices. If they set the price at $9000 they would be delighted, but the importance of this is muted by the fact that 95%+ of supply is contracted. And in overall terms, the high-price excursions that have occurred have still left average prices low. Some are concerned that generators have lost a great deal of money in some deregulated markets like Australia. But some ﬁrms have thrived in the electricity generation market; and if none have this, it indicates a need for a rise in the risk premium and the price, and the market will itself correct for the inadequate proﬁt as long as there is open entry. Occasional high-price excursions are important in reinforcing the need to contract. In Australia, there are some requirements on generators to explain their re-bids but they are cursory and really there to prevent a maverick generator purposefully trying to undermine conﬁdence in the bidding program by constant changes aimed at pure deception. As in all markets the insurance against “abuse” is competition. If there are pockets where high prices can be manipulated, this has its own proﬁt-oriented correction factor unless it is government induced, in which case there is a more straightforward deregulatory solution. Equally important to a competitive generation supply is a competitive retail market. Retailers in the electricity market are always likely to confront consumers requiring the product at will and at a known price. With smart metering, some price-induced demand shaving will be increasingly possible at the household level but quantities becoming available are always likely to disappoint – after all, smart metering and controls have been in place with large users but seldom activated. Electricity is not sufﬁciently costly to trigger major changes in behavior, and the suggested elasticities (-0.2) would probably not be achieved in the short term even if full knowledge of costs were available. For the retailer, the main game is likely to remain forecasting his customers’ demand and arranging for supplies to be made available through an array of contractual mechanisms. Not only is the retailer a crucial link in bringing together supply and consumer demand but the retailer’s exposure to price volatility forces it to adopt very prudent contracting

EBL

Resource Adequacy and Efﬁcient Infrastructure Investment

415

strategies. A retailer going bare and relying on the spot is engaged in extremely risky business since in the price ramping-up process prior to a blackout caused by insufﬁcient capacity, that retailer will go broke. Not only does this concentrate the mind of the retailer itself, but its creditors add a further discipline. The retailer’s creditors, conscious of their own exposure, are constantly viewing its books and ensuring their interests are protected as conditions of maintaining and extending loans. The consequent risk-aversion of competitive retailers is one reason why there need be little concern about the emergence of “gentailers” because retail arms would not favor their generator afﬁliates. Should they do so, they would be to jeopardize supplies from other generators. The more signiﬁcant concerns are about the interface of transmission and generation where one is market-provided and the other is centrally determined. No market has yet devised a practical means of marrying the two components of supply in a market-driven context. Australia’s Electricity Code, in principle, requires new generators to ensure that they have adequate transmission but, in practise, transmission remains regulated. We have outlined a means of moving to a market-oriented solution. References Australian Energy Market Commission (2006). Annual Electricity Market Performance Review: Reliability and Security, available at: http://www.aemc.gov.au/pdfs/reviews/Annual%20Electricity%20 Market%20Performance%20Review%20%20Reliability%20and%20Security%202006%20Report/ aemcdocs/001Draft%20Report.pdf. AGL (2005). Electricity Transmission Revenue and Pricing Rules Review: Issues paper on transmission revenue requirements, October, available at: http://www.aemc.gov.au/pdfs/reviews/Review%20of %20electricity%20transmission%20revenue%20and%20pricing %20Rules/Issues%20paper%20submissions/Issues%20Paper%20Transmission/000AGL.PDF. Boiteux, M.P. (1949). La tariﬁcation des demandes en pointe: Application de la théorie de la vente au coût marginal. Revue générale de l’électricité, 58, 321–40. Translated as “Peak load pricing”. Journal of Business 33, 157–79. Bushnell, J.B. and Wolfram, C. (2005). Ownership change, incentives and plant efﬁciency: The divestiture of U.S. electric generation plants., Center for Study of Energy Markets, available at: http://repositories.cdlib.org/cgi/viewcontent.cgi?article=1043&context=ucei/csem California Public Utilities Commission (2005) Capacity Markets White Paper, available at: http://www.ksg.harvard.edu/hepg/Papers/California. PUC. Capacity. Markets. White. Paper.pdf. Caramanis, M.C. (1982). Investment decisions and long-term planning under electricity spot pricing. IEEE Trans. Pow. App. and Sys., 101, 4640–8. Chao, H-P. (1983). Peak load pricing and capacity planning with demand and supply uncertainty. Bell J. Econ., 14, 179–90. Cramton, P. and Stoft, S. (2006). The Convergence of Market Designs for Adequate Generating Capacity with Special Attention to the CAISO’s Resource Adequacy Problem, A White Paper for the Electricity Oversight Board, p. 30, available at: http://www.ksg.harvard.edu/hepg/Papers/Cramton_ Stoft_0406.pdf. Fabrizio, K. M., Rose, N.L., and Wolfram, C. (2004). Does competition reduce costs? assessing the impact of regulatory restructuring on U.S. electric generation efﬁciency. Center for Study of Energy Markets, available at: http://www.ucei.berkeley.edu/PDF/csemwp135R.pdf. Grey, P., Lewis, P., and Grifﬁn, J. (2005). Peace/Vaasaemg: Utility customer switching research report. World Retail Energy Market Rankings. Haydon, J.J. and Michaels R.J. (2006). Merchant transmission redux. Pub. Util. Fort., September, 58–61. Hogan, W.W. (1988). Transmission investment and competitive electricity markets. Center for Business and Government, John F Kennedy School of Government, Harvard University, April. Industry Commission (1991). Report into Electricity.

EBL

416

Competitive Electricity Markets

Institute of Public Affairs/Tasman Institute (1991). Report on Victorian Electricity Corporatisation and Privatisation. Joskow, P. (2006). Markets for power in the United States: An interim assessment. The Energ. J., 27(1), 1–36. available at: http://econ-www.mit.edu/faculty/download_pdf.php?id=1219. KEMA Consulting (2005). Review of Methodology and Assumptions Used in NEMMCO 2003/04 Minimum Reserve Level Assessment. NEMMCO, available at: http://www.nemmco.com.au/ powersystemops/240-0009.pdf. Littlechild, S. (2005). Beyond regulation. Beesley Lectures on Regulation, Institute of Economic Affairs, London, available at: http://www.iea.org.uk/ﬁles/upld-article94pdf?.pdf. London Economics (1999). Report for the ACCC, April. Michaels, R.J. (2006). Vertical integration and the restructuring of the U.S. electricity industry. Pol. Anal., 572, 20. National Audit Ofﬁce (2003). The New Electricity Trading Arrangements in England and Wales, available at: http://www.nao.org.uk/publications/nao_reports/02-03/0203624.pdf. National Energy Policy Development Group (2001). Reliable, Affordable, and Environmentally Sound Energy for America’s Future, available at http://www.whitehouse.gov/energy/National-EnergyPolicy.pdf. Newbery, D.M. and Pollitt, M.G. (1997). The restructuring and privatisation of the CEGB – was it worth it?. J. Ind. Econ., 45(3), 269–303. Oren, S.S. (2000). Capacity payments and supply adequacy in competitive electricity markets. VII Symposium of Specialists in Electric Operational and Expansion Planning, Curitiba, Brazil, 21–26 May. Available at: http://www.pserc.wisc.edu/ecow/get/publicatio/2000public/oren_capacity_payment.pdf. Report by the Independent Committee of Inquiry (1993) (the “Hilmer Report”) National Competition Policy. Roques, F.A., Newbery, D.M., and Nuttall, W.J. (2005). Investment incentives and electricity market design: The British experience. Rev. of Net. Econ., 4(2), 93–128. Simshauser, P. (2006). The conditions necessary for the microeconomic reform of a power, generation industry. Econ. Pol. and Anal., 34(2). Stern, N., Peters, S., Bakhshi, V., et al. (2006). Stern Review: The economics of climate change, HM Treasury, 239. Stoft, S.E. (2002). Power System Economics: Designing Markets for Electricity. Piscataway (NJ): IEEE Press. Victorian Energy Networks Corporation (2002). Value of Customer Reliability Report.

EBL

Part IV Market Design Issues

EBL

EBL

This page intentionally left blank

Chapter 12 Promoting Electricity from Renewable Energy Sources – Lessons Learned from the EU, United States, and Japan REINHARD HAAS,1 NIELS I. MEYER,2 ANNE HELD,3 DOMINIQUE FINON,4 ARTURO LORENZONI,5 RYAN WISER,6 AND KEN-ICHIRO NISHIO7 1

Energy Economics Group, Vienna University of Technology, Vienna, Austria; BYG, Technical University of Denmark, Lyngby, Denmark; 3 Fraunhofer Institute for Systems and Innovation Research, Karlsruhe, Germany; 4 CIRED (Centre International de Recherche sur l’Environnement et le Développement), Nogent sur Marne, France; 5 IEFE, Bocconi University, Milan, Italy; 6 Lawrence Berkeley National Lab, Berkeley, CA, USA; 7 Central Research Institute of Electric Power Industry, Tokyo, Japan 2

EBL

Summary The promotion of electricity generated from renewable energy sources (RES) has recently gained high priority in the energy policy strategies of many countries in response to concerns about global climate change, energy security, and other reasons. This chapter compares and contrasts the experience of a number of countries in Europe, states in the United States as well as Japan in promoting RES, identifying what appear to be the most successful policy measures. 12.1. Introduction The current high standard of living enjoyed in industrialized countries owes much to the high per capita consumption of energy, an increasing portion of which comes from electricity. Yet, many experts believe that the current patterns of generating electricity, mainly from fossil and nuclear resources, are not sustainable in the long term. Moreover, as Ford explains in Chapter 14 in this volume, there are increasing concerns about the environmental costs associated with electricity generation, notably greenhouse gas emissions. For these and other reasons, some experts are convinced that we must ﬁnd a way to gradually switch to more sustainable energy conversion and energy use over time. Such a conversion is not easy for many energy-intensive applications such as air transportation, but is relatively less painful in case of electricity generation. 419

420

Competitive Electricity Markets

A number of studies have concluded that with efﬁcient use of energy, even a high standard of living can be sustained with renewable energy resources alone (e.g., Flavin and Lenssen, 1994; Scheer, 2001). While this may be an extreme case, many are convinced that we can – and should – rely on renewable energy sources (RES) for an increasing portion of our energy needs. In a 2005 study, the European Renewable Energy Council (EREC), for example, concluded as much as 40% of our energy needs could be supplied from RES by 2040. Theoretically at least, there is plenty of RES to go around. The amount of solar radiation falling on earth every day exceeds the energy we consume worldwide in a year. The problem, of course, is that it is widely dispersed and intermittent. But with human ingenuity and improved technology, more RES can be captured and put to use. Historically, the largest contributor has been hydro power resources and biomass. Technically speaking, the resource with the largest additional future potential1 for generating electricity from RES is wind energy, followed by photovoltaics (PV), solid biomass, hydro power, and biogas. Other options with vast potential include tidal and wave power as well as solar thermal electricity. The appeal of renewables continues to grow (e.g., Meyer, 2003) due to •

their contribution in reducing greenhouse gas emissions associated with current electricity generation; • reduced dependence on imported energy resulting in energy security and a more diversiﬁed resource base; • contribution to increases in local employment and income; and • working as a hedge against volatile fossil fuel prices, as well as avoiding risks of disruption in fossil fuel supplies.

EBL

Despite these advantages, renewables face a number of barriers if they are to contribute to the market on a large scale (e.g., Komor, 2003). The major barrier is the current cost disadvantage compared to electricity generated from fossil or nuclear fuels.2 A signiﬁcant switch to a renewable energy system would initially require substantial investments in infrastructure and then technical innovations. Yet, such a future is, at least in principle, not far-fetched because the costs of RES have been steadily declining and are likely to fall even faster as a result of learning-by-doing, the economies of scale, and expected technological progress. Moreover, one convincing argument for supporting RES is that other energy technologies have traditionally received – and continue to receive – enormous subsidies from governments (e.g., Oosterhuis, 2001). Aside from technological and investment obstacles, there are institutional, political, and legislative barriers as well as problems arising from lack of sufﬁcient grid capacity and public and political awareness in many countries. Additional barrier includes lack of adequate recognition and support in regulations,3 which limits the contribution of RES. To overcome these barriers, many governments have set ambitious targets and goals to promote electricity generation from RES in recent years. The European Union (EU), for 1 There are a variety of deﬁnitions of the renewable energy potential, e.g., Voogt et al. (2001) speciﬁes as follows: “theoretical potential” > “technical potential” > “realistic potential” > “realizable potential.” Note that the technical potential is substantially larger than the realizable potential that takes into account current non-technological factors. 2 This cost barrier is partly due to the lack of a level playing ﬁeld as long as externalities from fossil fuels and nuclear are far from being included in the consumer price. 3 See also Chapter 13 on distributed generation where the implications of absorbing large amounts of RES in the network are discussed.

Promoting Electricity from Renewable Energy Sources

421

example, issued a directive in 2001 with the target to increase the share of electricity from RES from 12% in 1997 to 21% by 2010 (EC, 2001). A decision made in 2007 sets an even more ambitious 20% RES target for total energy by 2020. Currently, a number of schemes are being implemented in different countries to increase the share of renewables in the energy mix. While the speciﬁcs vary, most schemes are attempting to • • • •

enhance social acceptance and increase public awareness of renewable energy; improve reliability, technical performance, and standardization; remove obstacles to grid connection; reduce administration and transaction costs while minimizing the ﬁnancial subsidies; and • ensure sustainable growth of the renewable energy industry. The major promotional strategies include investment subsidies, feed-in tariffs (FIT), tax incentives, portfolio standards, quota-based tradable green certiﬁcates (TGC), and tendering systems. While each scheme has certain advantages, there is no consensus on what may deliver the best results at the lowest cost. This is a crucial issue, especially if we are going to rely on renewables for a growing percentage of energy mix in the future. This chapter examines the experience gained from various regulatory and support strategies for the promotion of electricity generation from renewables. A secondary objective is to provide evidence to improve future policies. The chapter’s primary focus is on countries with considerable experience including the EU, the United States, and Japan. The chapter is organized as follows. Section 12.2 covers historical development and the future potential of renewables. Section 12.3 classiﬁes different types of promotional strategies and summarizes the experience up to 2005. Section 12.4 evaluates the most important promotional programs in different countries, and Section 12.5 discusses the relative merits of different strategies. The chapter’s main conclusions appear at the end.

EBL

12.2. Historical Overview For millennia until the advent of industrial evolution, humans relied on renewable energy for most of their needs, albeit at a mere fraction of what we typically use today. Over the past two centuries, humankind has increasingly relied on fossil fuels, which are blamed for global climate change. Figure 12.1 shows the recent pattern of electricity generation from renewables for EU-25, United States, and Japan for the period 1990–2004, where EU has managed a gradually increasing trend compared to the United States and Japan. The side bar in Fig. 12.1 shows the mix of renewables with the dominance of hydro power electric generation everywhere.4 An entirely new picture emerges if hydro power is excluded (Fig. 12.2). While in the early 1990s the United States had the highest non-hydro renewable generation, the EU has assumed this role since 2002, mostly due to a rapid growth of wind energy. The development in Japan can be characterized by modest growth in the waste-to-energy and geothermal power in 1990s and in PV and wind power in 2000s, with biomass resources such as black-liquor utilized in the pulp and paper industry playing an important role over the decades. 4

Annual ﬂuctuations in meteorological conditions explain the year-to-year ﬂuctuations.

Competitive Electricity Markets 500

100%

450

90%

Uunited States

80%

400

70%

350 EU-25

300

60%

250

50%

200

40% 30%

150

Japan

95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04

EU-25 United States

19

19

19

19

19

19

93

0%

94

10%

0 92

20%

50 91

100

90

Electricity generation (TWh/year)

422

Hydro power Biomass Geothermal

Japan Wind Waste Solar

Fig. 12.1. Recent pattern of electricity generation from RES in EU-25, the United States, and Japan (left) and breakdown of electricity generation from RES in 2004 (right). Sources: EUROSTAT (2007), IEA (2006b), METI (2007), Black and Veatch (2006, personal communication with Ryan Wiser).

100%

Electricity generation [TWh/year]

140

EBL

120 100

United States

80 60

80%

60%

40%

40 EU-25 Japan

20 0

97 19 98 19 99 20 00 20 01 20 02 20 03 20 04

EU-25

19

96

95

19

19

94

19

93

92

19

19

19

91

0%

90 19

20%

Wind Waste Solar

United Japan States Biomass Geothermal

Fig. 12.2. Historical pattern of electricity from non-hydro RES in EU-25, the United States, and Japan (left) and breakdown of the mix for 2004 (right). Sources: EUROSTAT (2007), IEA (2006b), METI (2007), Black and Veatch (2006, personal communication with Ryan Wiser).

Despite the recent gains in Europe, the recent pattern of growth of electricity generation from renewables is far from a success story. Between 1997 and 2004, for example, no country or region has been able to signiﬁcantly increase the share of renewables as a percentage of total electricity consumed as illustrated in Fig. 12.3. In the EU, the gain is a modest 2% from 12 to 14%, Japan shows a slight decrease while there has been a signiﬁcant drop in the United States, from about 14 to 9%.

RES-E share in gross electricity consumption

Promoting Electricity from Renewable Energy Sources

80%

80%

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

423

0% AT BE DK FI FR DE GR IE IT LU NL PT ES SE UK CY CZ EE HU LA LT MT PL SK SI

1997

2004

EU target by 2010

EU-25

United Japan States

Fig. 12.3. Share of RES including large hydro power in gross electricity consumption in EU-25, the United States, and Japan. For the EU-25, 2010 targets are shown. Sources: Ragwitz et al. (2007), IEA (2006b).

Within the EU, only a few countries such as Denmark, Finland, Hungary, and Germany have managed to increase their renewable shares considerably during the period and may be considered to be on target to meet the indicative targets as set in Directive 2001/77/EC.5 For the EU as a whole, the 2006 actual is far from the target set for 2010. For the United States, no speciﬁc target currently exists on the national level.6 Elsewhere, renewables continue to grow in absolute terms, but the story is pretty much the same: as a percentage of total electricity generation, renewables have a hard time keeping their current penetration levels.

EBL

12.3. Promotional Strategies With the exception of hydro power, some biomass and wind energy plants at favorable locations, most renewable energy technologies currently are at a cost disadvantage relative to conventional technologies. As already mentioned, part of this cost disadvantage is due to the fact that most conventional technologies have traditionally received – and continue to do so – signiﬁcant direct and indirect subsidies including those offered to nuclear energy and oil and gas exploration. Moreover, until recently, the full effect of the externalities, notably emissions of greenhouse gases, has not been reﬂected in prices. Renewable energy advocates argue that RES deserve similar subsidies to overcome their current cost disadvantage, pointing out that over time, these subsidies can be reduced or eliminated. Such arguments aside, the pace of development of electricity generation 5 In addition to these sector-speciﬁc targets, a 20% target on primary energy for all sectors except transport by 2020 was introduced in early 2007. Yet, no speciﬁc targets for electricity have been speciﬁed, which is a weakness of the new strategy. 6 However, in the United States, a number of states have set targets and currently there is a proposal for a 15% mandate by 2020.

Competitive Electricity Markets

424

from RES is closely linked to the level of ﬁnancial incentives and/or regulatory mandates, both of which are dependent on political decisions. And these are the main variables that differentiate the development pace and penetration of RES in different countries. This section examines a number of schemes for supporting renewables and compares the results. Table 12.1 provides a classiﬁcation of promotion strategies (Haas et al., 2004; Menanteau et al., 2003). Fundamentally, there are four basic ways to subsidize or promote RES: 1. Regulatory price-driven strategies. Under these schemes no quantity goals or targets are established. Instead, the focus is on providing generators with ﬁnancial support in terms of a subsidy per kW of capacity installed or a payment per kWh of energy produced. There are a number of variations under this scheme such as •

investment-focused strategies where ﬁnancial support is provided through investment subsidies, soft loans, or tax credits, usually per unit of generating capacity installed and • generation-based strategies where ﬁnancial support is offered as a ﬁxed payment or as a premium per unit of energy generated. Under a ﬁxed payment scheme such as FITs, generators receive a ﬁxed amount per kWh generated regardless of the costs of generation or price while under a premium scheme a ﬁxed amount is added to the electricity price. In practice, this makes a difference for the renewable plant owner. In the latter case, the total price received per kWh (electricity price plus the premium) is less predictable than under the FIT because it depends on a volatile electricity price. In principle, a mechanism based on a ﬁxed premium – one that reﬂects an environmental bonus for RES and penalizes conventional energy for their externality costs – could establish a level playing ﬁeld allowing fair competition between RES and conventional power sources.

EBL

Table 12.1. Promotional strategies for supporting RES (Haas et al., 2004; Menanteau et al., 2003). Direct

Regulatory

Voluntary

Price driven

Quantity driven

Investment focused

Investment subsidies Tax credits Low interest/soft loans

Tendering system for investment grant

Generation based

(Fixed) Feed-in tariffs Fixed premium system

Tendering system for long-term contracts Tradable green certiﬁcate system

Investment focused

Shareholder programs Contribution programs Green tariffs

Generation based

Indirect

Environmental taxes Simpliﬁcation of authorization procedures Connection charges, balancing costs

Voluntary agreements

Promoting Electricity from Renewable Energy Sources

425

Such schemes have the advantage of allowing renewables to penetrate the market quickly as their production costs drop below the electricity-price-plus-premium. Together with other incentives and considerations which tax conventional power sources in accordance with their environmental impact, well-designed ﬁxed premium schemes are theoretically one of the most effective ways of promoting electricity from RES. 2. Regulatory quantity-driven strategies. Under these schemes, the policymakers set a desired quota or goal, usually with a target date, to encourage the market penetration of RES. Examples include the following: •

Tendering or bidding schemes which call for tenders to acquire speciﬁc amounts of capacity or generation from speciﬁed types of RES. Competition between bidders leads to the winners of contracts which will receive a guaranteed tariff for a speciﬁed period of time. • Tradable certiﬁcate schemes such as renewable portfolio standards (RPS), which are popular in the United States, or TGC in Europe. These schemes typically obligate one or more parties involved in the electricity supply chain such as the generators, wholesalers, distribution companies, or retailers to acquire a certain percentage of electricity from RES in their energy mix. Most schemes allow parties to trade certiﬁcates to demonstrate compliance. Certiﬁcates can be obtained in three ways: – from their own renewable electricity generation; – by purchasing renewable electricity and associated certiﬁcates from other generators; and – by purchasing certiﬁcates without purchasing the actual power from a generator or broker.

EBL

The price of the certiﬁcates is determined by a market for certiﬁcates, such as in Nord Pool. 3. Voluntary approaches. Voluntary schemes are based on the willingness of consumers to pay a premium for green energy. There are two main categories: 1. Investment based schemes driven where individuals voluntarily contribute to renewable energy by providing up-front capital and 2. Generation-based schemes where consumers pay a volumetric premium for renewable electricity deliveries. 4. Indirect strategies. Aside from the schemes already mentioned, there are other strategies which may have an indirect impact on the dissemination of renewables, including •

various forms of eco-taxes on electricity produced from non-renewable sources such as carbon taxes, sulfur taxes, or other7 ; • CO2 emission allowances which are the subject of much talk 8 ; and • removal of subsidies previously given to fossil and nuclear generation. Indirect schemes could also include regulatory and institutional assistance including preferential permitting and siting, easy connection to the grid, and the operational concessions that make it easy to feed RES-generated power into the system. This is particularly important because most RES generation tends to be intermittent and unpredictable. 7

Promotion of renewable electricity via energy or environmental taxes can be achieved either by exempting RES from energy taxes or by providing a partial or total refund of any taxes collected. 8 Refer to Chapter 14 on global climate change for further discussion.

426

Competitive Electricity Markets

Preferential permitting and siting can reduce potential oppositions to RES plants if they address issues of concern, such as noise and visual or environmental impacts. Regulations can be used to set aside speciﬁc locations for development and/or to omit areas of higher risk of environmental damage or injury to birds. Standardization of interconnecting RES-generated power to the grid can ease requirements that are often overly burdensome and inconsistent and can lead to high transaction costs for RES project developers.9 Safety requirements are essential, in particular in the case of the interconnection in weak parts of the grid. However, unusually burdensome criteria on interconnections can lead to higher prices for access to the grid – or in cases are used as an excuse to deny access. Clarity, transparency, and reasonableness of safety and interconnection requirements are critical. Moreover the rules must be clear and fair for distribution of additional costs imposed by RES on the network. Finally, there must be clear rules delineating responsibility for physical balancing associated to intermittent production from some RES-E technologies, in particular wind power.10 12.3.1. Historical milestones The birth of today’s modern renewable energy industries may be traced largely to the pioneering efforts of private Danish investors and developers in the early 1970s and to the passage in the United States of the Public Utilities Regulatory Policy Act of 1978 (PURPA), which arguably introduced the earliest form of a mandatory feed-in law in the power industry. The state of California, for long a proponent of alternative forms of generation, developed an attractive and generous subsidy scheme – called standard contract for qualifying facilities or QFs – which, when combined with available federal and state tax credits, stimulated the deployment of renewable energy projects. Both PURPA and the California scheme, however, had their drawbacks because, arguably, they did not provide an adequate incentive for the deployment of efﬁcient technologies.11 Some critics have characterized the aggressive promotion of renewables in California as too much, too soon, and at too high a price.12 In contrast, the Danes implemented a testing and certiﬁcation procedure for wind turbines as early as 1978 as a pre-condition for receiving subsidies resulting in high reliability and productivity (Meyer, 2004a, b). In the early 1990s, promotional programs based on regulated and obligatory tariffs for the purchase of electricity from speciﬁed renewable sources became common and were reﬁned in various European countries. The most important schemes ﬁxed FIT and ﬁxed premium systems used in Denmark, Germany, and Spain to good effect. Under these schemes, utilities are legally obliged to pay the prescribed FIT as long as the RES plants meet certain technical standards. Not surprisingly, the 1990s saw a wind power boom in Europe – especially in Germany, Denmark, and Spain where generous FIT schemes were introduced. More than 80% of the European wind capacity installed at the end of the

EBL

9

Chapter 13 discusses regulatory aspects of encouraging distributed generation, RES, and combined heat and power (CHP). 10 These issues have become pronounced in countries like Germany and Denmark, with signiﬁcant penetration of wind. 11 Although the federal government plays an important role in providing tax incentives for renewables, states have historically been the innovators in supporting the commercial application of RES technologies in the United States. 12 This colorful phrase is attributed to Michael Peevey, a senior vice president at Southern California Edison at the time and currently the Chairman of the California Public Utilities Commission.

Promoting Electricity from Renewable Energy Sources

427

1990s was located in these three countries. The competitive tendering system, favored in the United Kingdom and France, has had limited success. Meanwhile, in the United States, after an initial growth spurt in the 1980s, the 1990s saw relatively little new development as the standard offers established in California and other states were no longer aggressively promoted. With the ongoing liberalization of electricity markets across Europe and other countries, TGC, based on quota obligations for RES, have become more prominent. In Europe this scheme has been tried in Italy, the United Kingdom, and Sweden in different variations, so far without great success. The ﬁrst application of such a quota-based system occurred in the United States, at the state level with or without TGC. Renewable energy quotas have recently become the most popular support scheme in the United States and an increasing number of states have implemented them. In general, whilst the main goal of early subsidies was to increase the supply security and fuel diversity, the focus of programs in recent years has shifted to reduce the emission of greenhouse gases. Table 12.2 summarizes the most important historical milestones for promotional strategies. Table 12.3 provides an overview of more recent promotion schemes to support electricity generation from RES in the countries investigated. In Europe, FITs serve as the main policy instrument, with the exception of Finland, which exclusively uses tax credits and investment incentives for the promotion of RES. Over time, many countries have gone through major changes and occasional reversals in their renewable support policies, either in response to what was or was not working, to make it more effective, or in response to political or public sentiment. Figure 12.4 shows the evolution of the main support instrument for selected countries over time. Countries where the FIT has been continuously in effect since 1997 are Austria13 , Germany, Greece, Luxembourg, Portugal, and Spain. Countries with major changes include:

EBL

•

• • • •

the United Kingdom and Ireland which replaced their problematic tender schemes by a quota regulation in combination with a TGC market in 2002 and 2005, respectively; a few others including Belgium and Italy who substituted their FIT systems for a quota regulation; Sweden which switched to a quota obligation complementing the already existing tax and investment incentives in some cases; the United States where tax credits have been complemented by the introduction of RPS obligations in a number of states starting in 2000; and Japan which introduced a quota system in 2003 on top of an existing voluntary net metering for PV that has been offered by utilities since 1992.

12.3.2. Institutional and political determinants of policy The wide variations in the choice of instruments, their timing, and intended goals may be largely explained by variations in political developments and intentions of different countries. Indeed, one can speculate a strong collation between the choice of instruments, the level of FIT or premium offered, and stability of the policies with the institutional and political environments in each country. Three main sets of parameters are at play in the institutional arena and further described in Box 12.1. 13

With a short interruption for small hydro power in 2001.

Table 12.2. Historical overview on promotion strategies for electricity generation from renewables (Haas et al., 2007). Year

Country

Type of strategy

Program name

1979–89 1978–89

DK US

Investment subsidies Feed-in tariffs/tax relief

PURPA

1989–96

DE

1991–93

DE

1990–99 1990–present 1992–94

UK DE AT

1992–97 1994–present 1991–96

IT US SE

1992–99 1992–99 1994–present

DK DE, CH, AT GR

Investment subsidies plus feed-in tariffs Investment subsidies plus feed-in tariffs Tendering system Feed-in tariffs Investment subsidies plus feed-in tariffs Feed-in tariffs Tax relief Investment subsidies/tax relief Feed-in tariff Feed-in tariffs Investment subsidies

1994–present

ES

1994–2005

JP

1992–present 1996–present

JP DE, CH, NL, AT, UK

Feed-in tariffs or ﬁxed premium systems Investment subsidies Voluntary net metering Voluntary green tariffs

Technologies addressed

“100/250 MW Wind Programm"

Wind, biogas All technologies (except large hydro power) Wind

“1000-Dächer-Programm”

PV

NFFO/SRO/NI-NFFO “Einspeisetarif” 200 kW PV-Program

Selected technologies PV, wind, biomass, small hydro power PV

EBL

“CIP 6/92” Tax production credits

“Kostendeckende Vergütung” 1994–now: “Operational Program for Energy and Competitiveness” “Royal Decree 2366/1994” resp. “Royal Decree 436/2004” “Residential PV System Dissemination Program” “Surplus electricity purchase menu" Various brands

All technologies Varies over time; focus on wind Wind, solar, biomass Wind, biomass PV PV, wind, biomass, small hydro power, geothermal All technologies (except large hydro power) PV PV Selected technologies

1996–present 1997–present

CH FI

Voluntary stock exchange Tax incentives

“Solarstrombörse” Energy tax

1998–present

DE

Labeled “green electricity”

1999–present 1999–2000

DE NL

2000–present 1998–present

DE, FIT US

2001–04 2002–present 2002–present

IT ES IT, UK, BE

Soft loans (Voluntary) green certiﬁcates Regulated rates Quota obligation with TGC/renewable energy funds Rebates FIT/premium Quota obligation with TGC

TÜV, Grüner Stromlabel e.V., Öko-Institut “100,000 Dächer-Programm"

2003–present

JP

Quota obligation with TGC

“Tetti fotovoltaici” All technologies All technologies (wave, waste, and large hydro power depend on the country) “Renewables Portfolio Standard”

2003–present 2003–present 2003–present

AT SE NL

2005–present

IT

FIT Quota obligation with TGC Mixed strategy (FITs, tax incentives, TGC) PV feed in

“Ökostromgesetz” All technologies. No waste MEP (Environmental Quality of Power Generation) “Conto Energia”

“Renewable energy act” “Renewables Portfolio Standards”“Clean Energy Funds”

EBL

PV Wind, mini hydro power (<1 MW), wood-based fuels PV, wind, biomass, small hydro power PV All technologies (exempt municipal waste incineration) Selected technologies Selected technologies

PV

PV, wind, biomass, small hydro power (≤1 MW), geothermal All technologies All technologies except hydro power and “non-pure” biomass PV

Table 12.3. Current promotion strategies for electricity from RES in major EU countries, the United States, and Japan. RES-E technologies considered Major strategy

Large hydro power

Small hydro power

Austria

FITs

No

Belgium

Quota/TGC + guaranteed electricity purchase

No

FITs + government grants FITs and premiums

No

Renewable Energy Act 2003. (Ökostromgesetz). Technology-speciﬁc FITs guaranteed for 13 years for plants which get all permissions between 1 January 2003 and 31 December 2004 and, hence, start operation by the end of 2006. Investment subsidies mainly on regional level. No decision yet on follow-up support after 2004 Federal: The Royal Decree of 10 July 2002 (operational from 1 July 2003) sets minimum prices (i.e., FITs) for RES-E.a On regional level promotion activities include: Wallonia: quota-based TGC system on electricity suppliers – increasing from 3% in 2003 up to 12% in 2010; Flanders: quota-based TGC system on electricity suppliers – increasing from 3% (no MSW) in 2004 up to 6% in 2010. Brussels region: no support scheme yet implemented Government grants for 30–55 % of investment. FITs are in place since January 2006 and are guaranteed for 15 years. FIT level in 2006: wind: 48–92E/MWh,b biomass, landﬁll, and sewage: 63E/MWh, PV up to 5 kW: 204E/MWh FITs in place since 2002. Adoption of the act in 2005 for plants installed after Jan 2006. Fixed tariff option and premium option are offered alternatively and are guaranteed for 15 years in the ﬁxed option. FIT level for 2006: wind: 85E/MWh ﬁxed and 70E/MWh premium, small hydro power: 81E/MWh ﬁxed and 49E/MWh premium, biomass/biogas: 77–103E/MWh ﬁxed and 44–69E/MWh premium, biomass coﬁring: 19–41E/MWh premium, geothermal: 156E/MWh ﬁxed and 126E/MWh premium, PV: 456E/MWh ﬁxed and 435E/MWh premium Act on Payment for Green Electricity (Act 478): ﬁxed (and low) premium prices instead of former high FITs for wind onshore. Tendering for offshore wind. Biomass and biogas receive FITs of 80E/MWh for the ﬁrst 10 years, 54E/MWh for the next 10 years. Net metering for PV used for individual houses

Cyprus

Czech Republic

Denmark

FITs and premiums, net metering for PV

“New” RES (wind on- and offshore, PV, solar thermal electricity, biomass, biogas, landﬁll gas, sewage gas, geothermal)

Municipal solid waste

FITs for waste with a high share of biomass

EBL

No

No

No

Subsidies available for CHP plants using waste

Estonia

FITs

No

Finland

Tax exemption

No

France

FITs

No

Purchase obligation of RES-E only valid for amount of network losses. Level of FIT since 2003: 52E/MWh for all RES guaranteed for 12 years. Program will expire in 2015 Tax refund: Mix of tax refund and investment subsidies: tax 4.2E/MWh (plant refund of 6.9E/MWh for wind and 4.2E/MWh <1 MW) for other RES-E. Investment subsidies up to 40% for wind and up to 30% for other RES-E Up to mid-2007 FITs for RES-E plant < 12 MW (wind plants are not due to the capacity limit) guaranteed for 15 years (20 years PV and hydro power). Tenders for plant >12 MW. After mid-2007 no limitation of capacity for FITs, provided that the equipment are in speciﬁc zones decided by local communities and regional administrations. (Energy act of July 2005). FITs in more detail: biomass: 49–61E/MWh, biogas and methanization: 75–90E/MWh, including premium for energy efﬁciency up to 120E/MW, including premium for “methanization” up to 140E/MWh, geothermal: 76–79E/MWh, PV: 300E/MWh (20 years) including premium (for “integration in buildings”) up to 550E/MWh, sewage and landﬁll gas: 45–60E/MWh, wind onshorec : 28–82E/MWh; wind offshored : 30–130E/MWh, hydro powere : 54.9–61E/MWh Novel of German Renewable Energy Act in 2004: FITs guaranteed for 20 years.f In more detail, FITs for new installations (2006) are: hydro power: 66.5–96.7E/MWh (30 years); windg : 52.8–83.6E/MWh; biomass and biogas: 81.5–171.6E/MWh; landﬁll, sewage, and landﬁll gas: 64.5–74.4E/MWh; PV: 406–568E/MWh; geothermal: 71.6–150E/MWh

No

Tax refund (2.5E/MWh)

FIT: 45–50E/ MWh

EBL

Germany

FITs

Only refurbishment

No

(Continued)

Table 12.3. (Continued) RES-E technologies considered Major strategy

Large hydro power

Greece

FITs + investment subsidies

No

Hungary

FITs + soft loans

Ireland

FIT

Italy

Quota/TGC + FITs for PV

Latvia

Quota + FITs

Lithuania

FITs

Netherlands

No support system

Small hydro power

“New” RES (wind on- and offshore, PV, solar thermal electricity, biomass, biogas, landﬁll gas, sewage gas, geothermal)

Municipal solid waste

No FITs guaranteed for 12 years with the possible extension to 20 years. Tariff levelh : wind onshore, small hydro power, geothermal, biomass and biogas: 73E/MWh (mainland) and 84.6 E/MWh (islands); wind offshore: 90E/MWh (mainland and islands); solar thermal: 230–250E/MWh (mainland) and 250–270MWh (islands); PV: 400–450E/MWh (mainland) and 450–500E/MWh (islands) and investment incentives: 30–40% of investment incentives or 100% tax exemptions are offered by law 3299/2004. Reduction of taxable income of up to 20% FIT: 35– Technology-speciﬁc FITs since 2005. Tariff level (2006)i : geothermal, biomass, Waste: 69E/MWh biogas, small hydro power (<5 MW): 39–108E/MWh, solar and wind: 95E/MWh; 39–108E/ cogeneration: 36–69E/MWh. Soft loans from the Hungarian Development Bank MWh No FITs are granted for 15 years. Tariff level (2006): wind: 57–59E/MWh; landﬁll gas: No 70E/MWh; other biomass: 72E/MWh; small hydro power: 72E/MWh Quota obligation (TGC system) on electricity suppliers: 3.05% target (2006), increasing yearly 0.35% up to 2008; TGC issued for all (new) RES-E (incl. large hydro power and MSW) – with rolling redemptionj ; no penalty; 12 528E/MWh (2006) to purchase TGCs from the grid operator, but market distortions appear.k Feed in tariff for PV and locally investment subsidies from regional administrations The FIT scheme has been replaced with a yearly quota system in 2003, but some No RES-E producers continue receiving the FITs. However, political framework conditions for the support of RES-E are currently under development No FITs are in place since 2002. Tariff level: small hydro power: 57.9E/MWh; wind: No 63.7E/MWh; biomass: 57.9E/MWh. The implementation of a quota obligation is planned for 2021 No FIT scheme was abolished in summer 2006 since the government expects to fulﬁll No the 2010 target set by the EC without further ﬁnancial support and RES-E support costs were higher than expected

EBL

Poland

Quota/TGC

Portugal

FITs + investment subsidies

No

Slovak Republic

FITs and tax exemption

No

Slovenia

FITs or premiums

No

Poland applies a green power purchase obligation since 2003 and started the certiﬁcate trading in December 2005. The quota target to be fulﬁlled is set at 3.6% in 2006 and increases up to 7.5% in 2010 FITs (Decree Law 33-A/2005) and investment subsidies of roughly 40% (Measure 2.5 (MAPE) within Program for Economic Activities (POE)) for wind, PV, biomass, small hydro power and wave. Average FITs in 2006: windl : 74E/MWh; wave: n.a.; PVm : 310–450E/MWh; small hydro power: 75E/MWh FITs since 2005 (Decree No. 2/2005): wind: 75.1E/MWh; small hydro power (<5 MW): 61.7E/MWh; solar: 214.6E/MWh; geothermal: 93.9E/MWh; biomass: 72.4E/MWh. Tax exemption for RES-E in ﬁrst 5 years of operation FITs were introduced in 2004 and are granted for 10 years. There are two alternatives, the ﬁxed tariff option and a premium payment which is paid on top of the market price. Tariff levels (2004–present): small hydro power: 59–62E/MWh (ﬁxed) and 26–28E/MWh (premium); biomass: 68–70E/MWh (ﬁxed) and 34–36E/MWh (premium); landﬁll and sewage gas: 49–53E/MWh (ﬁxed); biogas: 121E/MWh (ﬁxed); wind: 59–61E/MWh (ﬁxed) and 25–27E/MWh (premium); geothermal: 59E/MWh (ﬁxed) and 25E/MWh (premium); solar: 65–374E/MWh (ﬁxed) and 31–341E/MWh (premium) FITs (Royal Decree 436/2004): RES-E producers have the right to opt for a ﬁxed FIT or for a premium tariff.o Both are adjusted by the government according to the variation in the average electricity sale price. In more detail (2006): wind, biomass, small hydro power (<25 MW), geothermal: 68.9E/MWh (ﬁxed) and 38.3E/MWh (premium); solar thermal and PVp : 229.8–440.4E/MWh, 194E/MWh;agricultural and forest residues: 61.3E/MWh (ﬁxed) and 30.6 (premium). Moreover, soft loans and tax incentives (according to “Plan de Fomento de las Energías Renovables”) and investment subsidies on regional level

FIT for urban waste: 75E/MWh No

No

EBL

Spain

FITs or ﬁxed premiums

Depending on the plant sizen

FIT: 53.6E/MWh (ﬁxed) or 23E/MWh (premium)

(Continued)

Table 12.3. (Continued) RES-E technologies considered Major strategy

Large hydro power

Small hydro power

“New” RES (wind on- and offshore, PV, solar thermal electricity, biomass, biogas, landﬁll gas, sewage gas, geothermal)

Sweden

Quota/TGC

No

United Kingdom

Quota/TGC

No

Quota obligation (TGC system) on consumers: increasing from 7.4% in 2003 up to 16.9% in 2010. For wind investment subsidies of 15% and additional small premium FITs (“Environmental Bonus“q ) are available Quota obligation (TGC system) for all RES-E: increasing from 3% in 2003 up to 10.4% by 2010 – buy-out price is set at 32.33£/MWh for 2005/2006. In addition to the TGC system, eligible RES-E are exempt from the Climate Change Levy certiﬁed by Levy Exemption Certiﬁcates (LECs), which cannot be separately traded from physical electricity. The current levy rate is 4.3£/MWh. Investment grants in the frame of different programs (e.g., Clear Skies Scheme, DTI’s Offshore Wind Capital Grant Scheme, the Energy Crops Scheme, Major PV Demo Program, and the Scottish Community Renewable Initiative) Quota obligations established in 21 US states – covering 40% of US load – by 2006. Design and level of quotas varies by states. 10-year, 1.9 cent/kWh federal production tax credit available to certain “new” RES-E sources. 30% federal investment tax credit available for solar installations. 15 states have developed renewable energy funds, typically collected through electricity surcharges, and these funds have developed incentive programs that spend roughly $500 million per year on various RES-E technologies, through a variety of different kinds of programs. The largest of these programs involves rebates and production incentives for customer-sited photovoltaic systems in California

EBL

United States

Quota/TGC + tax incentives + state funds

No

Municipal solid waste

No

No

Japan

Quota/TGC + tax incentives + voluntary net metering

No

Quota obligation (TGC system): increasing from 3.3 TWh in 2003 up to 12.2 TWh (approx. 1.35%) in 2010, 16 TWh (approx. 1.63%) in 2014. TGC issued for all RES-E [exempt large hydro power (>1 MW), non-biomass fraction of MSW, conventional geothermal]. Maximum price of TGCs is set at 11JPY/kWh. Voluntary net metering for PV offered by power companies: the current purchase price is about 23JPY/kWh

Yes (exempt non-biomass fraction)

a FITs are guaranteed on national level for the ﬁrst 10 years of operation, e.g., in case of offshore wind in size of 90E/MWh. Note, they can only be claimed exclusively – in other words, they cannot be claimed if support is given by the regional TGC systems. b Stepped FIT: 92E/MWh for the ﬁrst 5 years of operation and then between 48 and 92E/MWh for the next 10 years. c Stepped FIT: 82E/MWh for the ﬁrst 10 years of operation and then between 28 and 82E/MWh for the next 5 years depending on the quality of site. d Stepped FIT: 130E/MWh for the ﬁrst 10 years of operation and then between 30 and 130E/MWh for the next 10 years depending on the quality of site. e Producers can choose between four different schemes. The ﬁgure shows the ﬂat rate option. Within other schemes tariffs vary over time (peak/base, etc.). f The law includes a dynamic reduction of the FITs: for biomass 1.5% per year, for PV 6.5% per year, for wind 2%, and for geothermal 1% per year. g Stepped FIT: in case of onshore wind 83.6E/MWh for the ﬁrst 5 years of operation and then between 52.8 and 83.6E/MWh depending on the quality of site. h Depending on location (islands or mainland). i Tariffs are differentiated depending on load distribution. j In general only plant put in operation after 1 April 1999 are allowed to receive TGCs for their produced green electricity. Moreover, this allowance is limited to the ﬁrst 8 years of operation (rolling redemption). k GSE (Italian Agency for renewable support schemes) inﬂuences strongly the certiﬁcates market selling its own certiﬁcates at a regulated price – namely at a price set by law as the average of the prices paid to acquire electricity from RES-E plant under the former FIT program (CIP6) minus the income from the sale of such electricity in the market. l Stepped FIT depending on the quality of the site. m Depending on the size: <5 kW: 420E/MWh or >5 kW: 224E/MWh. n Hydro power plants with a size between 10 and 25 MW receive a tariff of 68.9E/MWh (ﬁxed) or 38.3E/MWh, larger plants (25–50 MW) can opt for a ﬁxed tariff of 61.3E/MWh or a premium payment of 30.6E/MWh. o In case of a premium tariff, RES-E generators earn in addition (compared to ﬁxed rate lower) to the premium tariff the revenues from the selling of their electricity in the power market. p In case of PV the expressed premium tariff refers to plant > 100 KW. For small-scale plants (<100 kW) only a ﬁxed FIT is applied. q Decreasing gradually down to zero in 2009.

EBL

436

Competitive Electricity Markets

EBL

Fig. 12.4. Evolution of the main support schemes in EU-15 member states and the United States14 (based on Ragwitz et al., 2007; Wiser et al., 2007).

1. The political, culture, and policy style: Countries with a market-oriented political culture and a liberalistic government tend to choose the most market-based instrument with belief in the efﬁciency of market incentives and the desire to make the RES promotion instrument compatible with the electricity market principles. 14 There are also some limited FITs in the United States, but these are not too signiﬁcant and are not indicated in this table.

Promoting Electricity from Renewable Energy Sources

437

2. The convergence of supply security and CO emissions reduction: Countries with meager domestic energy supply and/or little fuel diversity tend to be concerned with energy security and may be concerned about CO emissions. 3. The absence of a strong conventional energy equipment-related industry: Countries who lack major manufacturing capabilities in fossil or nuclear power sector tend to be strong proponents of a national RES industry.

Box 12.1 Why European countries differ in their RES promotion policies •

•

•

•

•

•

United Kingdom: The country’s energy sector is focused on market forces and competition as the main means of controlling costs. This explains the government’s early reliance on tending process with cost cap, and then TGC obligations with buy-out provisions, at the expense of effectiveness of the instrument. United Kingdom currently enjoys a high degree of energy selfsufﬁciency in conventional energy resources and is less concerned about supply security. Moreover, Britain has been able to dramatically reduce emissions by replacing coal generation with natural gas. Germany: The promotion of RES is primarily driven by the politics of CO2 reduction, the phase-out of nuclear energy, and an increased dependence on imported energy. This explains why Germany was among the ﬁrst two countries to introduce generous and stable FIT instruments. Denmark: Since 1990s the Danish government has given high priority to RES and energy conservation for environmental reasons. Wind power has traditionally been promoted through generous schemes. Another motivating factor has been a fast-growing domestic wind power industry. The Danish strategy has changed after a shift of government in 2001. The incoming liberalconservative government is relying mainly on market forces to promote RES. As a result no net increase in land-based wind power has taken place since 2003. Spain: Increasing energy dependence and CO emissions are major issues. Spanish government and industries feel constrained by a nuclear moratorium. The regional and local communities contribute to the momentum by helping to facilitate the planning and location process. Italy: Cogeneration and renewables have been promoted since the 1990s under a FIT system inspired by the PURPA experience in the United States and to alleviate investment constraints. Renewables are also seen as contributing to reduce the country’s growing energy imports exasperated by the nuclear moratorium of the mid-1980s. Italy has adopted intensive market reforms and has introduced a TGC system in place of the FIT system to be consistent in its electricity market reform. France: Thanks to the powerful political lobby for nuclear energy, France relies on atom for roughly 70% of its electricity generation, complemented by hydro power production. Energy equipment manufacturing industry is quite successful in the exports markets. A renewables policy was developed to

EBL

Competitive Electricity Markets

438

maintain the nuclear option while giving a perception of support for renewables. Despite the adoption of a generous FIT system in 2001, the net effect has been modest to date due to planning and siting difﬁculties.

12.4. Comparison of Strategies This section summarizes the major national programs which have been implemented in different counties including a description of support system and policy targets and changes over time; a discussion of the attractiveness of the scheme from investors’ point of view; an overview of the effect of the policy; as well as an overview of the pros and con of alternatives. 12.4.1. Feed-in tariffs and premiums In Europe, FIT began to attract attention in the late 1980s especially in Denmark, Germany, and Italy followed by Spain in the 1990s. It is the most widely used promotion instrument in Europe. Figure 12.5 shows variations in FIT for electricity from onshore wind turbines between 2003 and 200515 indicating a broad range of support, some varying over time. In 2005, the schemes varied between 60 and 90 EUR/MWh due to differences in wind conditions and different levels of support in different countries. The FITs attract a lot of capacity, since a ﬁxed tariff is guaranteed, but only if the FIT is set at a level sufﬁcient to meet investors’ needs. This is evident in countries with substantial growth in wind power such as in Denmark in the 1990s and Germany and Spain since the late 1990s.

Level of fixed tariffs for wind onshore (€/MWh)

EBL

120 100 80 60 40 20 0 2003

2004

2005

Austria

France

Germany

Netherlands

Portugal

Spain – fixed tariff

Czech Republic

Slovenia

Fig. 12.5. FITs for electricity generated from onshore wind in selected European countries.

15

The performance of FIT is discussed in Section 12.5.

Promoting Electricity from Renewable Energy Sources

439

The experience of the United States, Germany, and Spain with FIT are highlighted below. 12.4.1.1. Implementation of PURPA in the United States The earliest form of a mandatory FIT may be traced to Public Utilities Regulatory Policy Act of 1978 (PURPA) in the United States. PURPA required utilities to purchase power from qualifying facilities (QF) including small renewable generators and combined heat and power (CHP) plants. Pricing varied by state but, especially in California, prices were tied to the price of the marginal conventional fossil fuel, which was high and projected to increase at the time, yielding highly attractive returns to renewable energy investors (Martinot et al., 2005). PURPA faced early legal challenges, but once it was underway, and where it was implemented aggressively, it enabled an environment in which renewable developers could secure ﬁnancing for their projects because they could sell their output under attractively priced long-term standard contracts. There was little risk under the scheme. PURPA was implemented differently in each state. The state of California developed a particularly generous standard contract, some as long as 15–30 years with a ﬁxed tariff for the ﬁrst 10 years of facility operation. Combined with favorable state and federal tax credits, the growth of QF capacity, which included renewables, was astounding, aided by California’s diverse and abundant renewable energy resources. Over a short period of time, about 12 000 MW of geothermal, small hydro power, biopower, solar thermal, and wind power capacities were constructed in the United States in the 1980s, of which more than half were in California (Martinot et al., 2005). But PURPA also had unintended negative consequences. By providing high proﬁtability, it created an over-heated renewable energy market in which project development arguably occurred at a pace that exceeded the ability of the industry to deliver efﬁciently. Moreover, because the incentives were capacity based, there was more of an incentive to deploy capacity (MW) rather than generate electricity (MWh). Finally, because of the generous incentives that were offered and the high number of project failures, PURPA-inspired QFs resulted in a backlash within the regulated utility industry, which was obligated to buy not only the power, but also within the ﬁnancial community and among some policymakers. This negative reaction arguably set the industry back to some degree (e.g., Martinot et al., 2005). Partly as a result of these factors, the US renewable energy market remained largely stagnant between the late 1980s and the year 2000 as state level implementation of PURPA became less aggressive and certain federal and state tax incentives were allowed to expire.

EBL

12.4.1.2. Implementation of FIT in Germany In Germany, a ﬁxed FIT scheme has been in place since 1991 when the Electricity Feed-in Act was passed. In 2000 this act was substituted by the Renewable Energy Act and a 12.5% target for the share of RES in electricity generation to be achieved by 2010 was established. The most important change has been the uncoupling of the tariff level from the electricity retail price and the setting of new tariffs based on the actual generation costs of a technology. This means that tariffs are differentiated not only by technology, but also within a given technology. Moreover, the tariff is adjusted according to location-speciﬁc generation costs inﬂuenced by wind speed, size of a plant, or the fuel type in case of biomass. Another feature of the Renewable Energy Act is a tariff degression for new installations designed to encourage technological development and learning. The act was amended

Competitive Electricity Markets

440

in 2004 with a 20% target for the share of renewables in electricity generation by 2020. The FIT for onshore wind has been reduced and sites with poor wind conditions have been excluded. Tariffs for geothermal electricity, small-scale biomass plants, and PV were increased. Furthermore, additional bonuses were granted for innovative technologies and refurbishment of large hydro power plants. Investment security for generators of green electricity is virtually guaranteed by FITs for a time scale of up to 20 years. As shown in Fig. 12.6, roughly two thirds of the increase in electricity generation from RES of about 40 TWh since the early 1990s may be attributed to onshore wind. Since 2000 – the year of the implementation of the Renewable Energy Act – more wind capacity has been connected to the German grid than all previous years. By 2005, Germany was getting about 11% of its electricity from RES in 2005, compared to about 4% in 1997. Despite the impressive gains, the German FIT scheme has its share of critics, mostly large incumbent utilities and energy-intensive customers who complain about the extra cost burden for the promotion of RES. Proponents of the FIT point out (Fig. 12.23) that the German scheme is only marginally more generous than others in Europe. Another problem is that wind power plants are concentrated in the northern part of Germany straining the local transmission network, which is also inﬂuencing load ﬂows in neighboring European countries. In balance, the German FIT may be considered a success story, albeit coming at a price. 12.4.1.3. FIT in Spain The dominant policy instrument for the promotion of electricity from renewables in Spain is a FIT, which has been in place since 1994.16 The same year, Spain established a 12% national target for renewables in total energy consumption by 2010. To further encourage investment in wind, the FIT scheme was amended in 2004 to effectively guarantee payments during the whole lifetime of a plant and additional incentives were introduced. Green electricity can be sold in the market either by using a bidding system or through

Electricity generation (GWh/year)

EBL

60 000 50 000 40 000

Wind onshore Hydro power small-scale Solid biomass PV

Hydro power large-scale Biogas Biowaste

30 000 20 000 10 000 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Fig. 12.6. Impact of German FIT on RES in Germany. Sources: IEA (International Energy Agency) (2006b), EUROSTAT (2007).

16

In 1998 two alternative payment options for green electricity generation were introduced, a ﬁxed tariff scheme and a premium tariff, which was paid on top of the electricity market price. The choice is valid for 1 year, after which the generator may decide to maintain the tariff option or change to the alternative option. Under both payment options, grid connection and purchase of the green electricity are guaranteed.

Electricity generation (GWh/year)

Promoting Electricity from Renewable Energy Sources

Wind onshore

Hydro power small-scale

Biogas

Solid biomass

Biowaste

PV

30 000

441

20 000

10 000

0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Fig. 12.7. Growth of electricity generation from renewables excluding large hydro power in Spain. Source: EUROSTAT (2007).

bilateral contracts. By the end of 2004, the overall remuneration level under the market option has increased more than was expected due to rising electricity market prices. As a result of these favorable conditions, the deployment of RES in Spain started to take off in the late 1990s with 30 TWh of additional generation in 2006, mostly from onshore wind (Fig. 12.7). The only criticism of the scheme is that the premiums offered may be too generous for wind generators. Yet, by and large, the Spanish scheme may be characterized as another European success story because it has resulted in a signiﬁcant increase in deployment of renewables with modest subsidies (Figs 12.5 and 12.12) in a relatively short period of time. Continuity and stability of the policy even under changing governments have contributed signiﬁcantly to the success of the policy.

EBL

12.4.2. Bidding/tendering systems Government tendering schemes to promote RES have been used in the 1990s in France (for wind energy and biomass), Ireland (The Irish Alternative Energy Requirement), Denmark (the last two offshore wind farms), and the United Kingdom, as well as in many states in the United States. The most well-known of these promotion strategies is the non-fossil fuel obligation (NFFO) in England and Wales, which is further described below. Similar schemes have been used in Scotland (Scottish Renewables Order, SRO) and Northern Ireland (NI-NFFO). However, in most cases, the schemes did not work effectively and starting in 2001/2002 the competitive tendering schemes were abandoned. Figure 12.8 compares the relative effectiveness of bidding vs FITs for wind energy in Europe prior to 2001 clearly showing the superiority of FIT schemes. Partly as a result of this, the United Kingdom switched to a renewables obligation scheme in 2002. More recently, both Ireland and France have also changed to FIT systems. 12.4.2.1. United Kingdom’s non-fossil fuel obligation (NFFO) As originally envisioned, the United Kingdom’s NFFO was to deliver 1500 MW of installed capacity from RES by the year 2000. The rational for the competitive tendering scheme was to invite developers to bid to construct renewable energy capacity. The tendering process would select among the viable proposals the least cost options within each technology grouping. To facilitate ﬁnancing, the winners would be awarded relatively long-term

Competitive Electricity Markets

442 350

300

W/capita

250

200

150

100

50

Fr

an

ce

K U

Ire

la

nd

n ai Sp

m er G

D

en

m

ar

an

k

y

0

Feed-in tariffs

Bidding

EBL

Fig. 12.8. Comparison of the relative effectiveness of FIT and competitive tendering schemes in promoting wind power deployment, 1990–2001. Source: Haas (2003).

contracts, up to 15 years with a guaranteed surcharge per unit of output for the entire contract period. The difference between the surcharge paid to NFFO generators (premium price) and a reference price (pool selling price) was to be ﬁnanced by a levy on all electricity sales of licensed electricity suppliers. The costs of this levy17 were to be passed on to consumers (Mitchell et al., 2005). In total, ﬁve tendering rounds were conducted in England and Wales resulting in 880 contracts being awarded. The competitive bidding resulted in declining prices over time as expected. Since the ﬁrst round in 1990, average prices have decreased from 6.5p/kWh to 2.71p/kWh (Fig. 12.9). Even lower prices, less than 2p/kWh, were obtained in Scotland for wind power, lower than electricity from coal, oil, nuclear, and some natural gas. The scheme provided revenue security as long as the plant operated. On surface, this appeared to be a successful scheme. However, things did not go smoothly in practice. Many of the awarded contracts did not materialize while others failed to meet the expected capacity targets (Fig. 12.10). Many factors contributed to this including submission of unrealistic bid prices to secure a contract and failure to obtain planning and other consents. Similar experiences with contract failure have been common in other bidding schemes (e.g., Wiser et al., 2006b). One lesson from this experience is that tendering schemes lacking penalties for non-delivery may be deﬁcient compared with other subsidy schemes. In 2002, the NFFO was replaced 17

The levy remains now only to continue the previously contracted arrangements.

Promoting Electricity from Renewable Energy Sources

443

8.0 7.0

cent / kWh

6.0 5.0 4.0 3.0 2.0 1.0 0.0 NFFO-1

NFFO-2

NFF O-3

1990

1991

1992

Average price (cent/kWh)

NFFO-3 1993

1994

1995

1996

Landfill gas

Wind

NFFO-4

NFFO-5

1997

1998

Hydro power

Fig. 12.9. Bid prices under NFFO in England and Wales, prices in EUR-cent/kWh. Source: www.ofgem.uk.

EBL

1400 1200

MW, MW/year

1000 800 600 400 200 0 NFFO-1

NFFO-2

NFFO-3

1990

1991

1992

NFFO-3 1993

Projected capacity (MW/period)

1994

1995

1996

NFFO-4

NFFO-5

1997

1998

Installed capacity (MW/period)

Installed capacity cum (MW)

Fig. 12.10. Capacities installed vs projected under NFFO in England and Wales. Source: www.ofgem.uk.

444

Competitive Electricity Markets

by the Renewables Obligation (RO; see Section 12.4.4) while contracts already awarded continue to be valid under the older terms. 12.4.3. Quota-based trading systems Alternative forms of quota-based systems are now in place in ﬁve EU countries, in some 21 states and the District of Columbia in the United States, and in Japan. The European quota systems are based on tradable generation certiﬁcates, TGC (Table 12.4) while in the United States, the schemes are referred to as RPS (Table 12.5). A comparison of recent TGC prices are presented in Fig. 12.11, showing relatively ﬂat price levels in most countries, over this 3-year period, but wide disparities in pricing across systems. The salient features of some of the schemes follow. 12.4.3.1. European TGC schemes Currently, quota-based TGC systems are in effect in the United Kingdom, Sweden, Italy, Belgium, and Poland. In the United Kingdom, Belgium, and Poland suppliers have to demonstrate compliance with the obligation; in Sweden the end-users are responsible while in Italy the quota has to be fulﬁlled by the producers and in a rather complicated way described below. In all cases the obligations can be met by • •

producing certiﬁcates by generating electricity from qualifying renewable plants; purchasing TGCs from other eligible generators, other suppliers, traders, or through organized exchanges; or • paying a penalty or “buy-out price” set by the regulatory authority.

EBL

In the United Kingdom the RO scheme came into effect in 2002, starting at 3.4% coverage of electricity demand for the period of 2003/2004, gradually increasing to 10.4% by 2010/2011, and remaining at that level until 2027. The major problems with the British RO scheme are that certiﬁcate prices are high, although slightly decreasing from 2003 to 2005 (Fig. 12.11), and that so far the quota has never been fulﬁlled. For example in 2004 only 2.2% of electricity has been generated from new RES vs the 3.4% speciﬁed in the quota.18 There are several explanations for this. In fact, not meeting the target is also a function of at least19 three major factors: (1) the low penalty, respectively, the fact that this penalty is recycled to the renewable generators (see above); (2) location and permitting constraints; and (3) banking is not allowed so RES generators fear (with good reasons) that the closer they come to the quota the lower will be the ROC price. Note that this is despite the fact that long-term contracts are possible and most of the certiﬁcate handling takes place within vertically integrated large companies. The penalty mechanism in the United Kingdom deserves special attention. All penalty payments are placed in a central fund. This fund is redistributed to suppliers which have met the obligation in proportion to the number of ROCs each supplier has presented. Therefore the real costs for a supplier who is not complying with the obligation are higher than their total buy-out price payments (“ﬁnes”). In contrast, accomplishing and surpassing 18

Notice that, because of multiple risks for the producers, developers, and obligated suppliers, most of the quotas are complied within long-term contracts between suppliers and producers and the exchange of certiﬁcates does not play the role that the theory could suggest. 19 Of course, more investigations are necessary to get detailed insight on the effects of a hybrid instrument (control by quantity – the quota – and by price –the buy-out price) to explain its poor effectiveness.

Table 12.4. Quota-based TGC systems in EU and Japan. United Kingdom

Belgium (Flanders)

Belgium (Walloon)

Italy

Poland

Sweden

Japan

Period Obligation

Start 2002 3% in 2003, 10.4% in 2010

Start 2002 1.2% (2003), 2% (2004) increasing up to 6% in 2010

Start 2001 2% in 2002 and increased annually by 0.35% between 2004 and 2008

Start 2005 7.5% in 2010

Start 2003 7.4% in 2003, 16.9% in 2010

Start 2003 Approximately 0.4% in 2003 Approximately 1.4% in 2010 Approximately 1.6% in 2014

Obligation on Technology bands (baskets) within overall quota Involved technologies

Supplier

Supplier

Start 2002 3% in 2003 increasing up to 12% in 2010 From September 2010 onward, the quota will be multiplied by 1.01 Supplier

Supplier

End-user

Supplier

No (introduction of technology banding is planned for the future) Small hydro power*, wind, biomass, solar-, geothermal energy, no waste

No

No

Producers and importers No

No

No

No

All renewables, no solid municipal waste

All renewables and high-quality CHP

All new renewables (incl. large hydro power, MSW, hydrogen and CHP)

Small and large hydro power, wind, biomass

PV, wind, biomass, small hydro power up to 1 MW, geothermal (exempt conventional type)

Existing plants eligible

No

Yes

Yes

No (for certiﬁcate issue), yes (for quota fulﬁllment)

No

Small hydro power (<1.5 MW), large hydro power (only some cases), wind, biomass, geothermal, wave Yes (small hydro power)

EBL

Yes

(Continued)

Table 12.4. (Continued) United Kingdom

Belgium (Flanders)

Belgium (Walloon)

Italy

Poland

Sweden

Japan

International trade allowed

No

No

No

No

Trading scheme with Norway planned but now abolished

No

Floor price

Not planned

Yes, but only in exchange with physical electricity and with countries that allow reciprocity Not planned

No

Floor prices for the introductory phase (in E/MWh):2003: 6.6; 2004: 5.5; 2005: 4.4; 2006: 3.3; 2007: 2.2;From 2008 onwards no ﬂoor price is planned

No

The buy-out price is 100 EUR/MWh

150% of the market price – but with a maximum of about 19E/MWh in 2004 and 26E/MWh in 2005

1 million JPY to non-fulﬁllment supplier

Power exchange

Open

Bilateral or in the market managed by private brokers

Penalty

The buy-out price is £30.51/MWh (for 2003/2004) (∼45E/MWh)

At federal level: from 1 July 2003 onward the grid operator is obliged to buy TGC issued anywhere in Belgium for the minimum prices per TGC (in size of 1 MWh) of:offshore wind 90E, onshore wind 50E, hydro power 50E, solar energy 150E, biomass 20E; within the Wallon region, RES-E producers may exchange their TGC for a subsidy at a ﬁxed price of 65E 75E/MWh From 1 April (in 2003; 2003 onward: 10E/MWh 100E/MWh in 2004; and (100E per 125E/MWh missing TGC in in 2005 size of 1 MWh)

Trading scheme

Stock exchange

Stock exchange

∗

Open, trading, and direct support

Hydro power plants with a capacity of less than 20 MW. Sources: Ragwitz et al. (2007), and METI (2007).

EBL No penalty is set; the grid operator sells certiﬁcates at a ﬁxed price 12 528E/MWh (2006) Bilateral or in the TGC market

Promoting Electricity from Renewable Energy Sources

447

Table 12.5. Current RPS schemes in the United States. State

Start date

Ultimate target

Existing plants eligible

Technology bands or tiers

Arizona California Colorado Connecticut Delaware Hawaii Iowa Maine Maryland Massachusetts Minnesota Montana Nevada New Jersey New Mexico New York Pennsylvania Rhode Island Texas Washington Washington DC Wisconsin

2001 2003 2007 2000 2007 2010 1999 2000 2006 2003 2005 2008 2003 2001 2006 2006 2007 2007 2002 2012 2007 2001

15% (2025) 20% (2010) 10% (2015) 10% (2010) 10% (2019) 20% (2020) ∼2% (1999) 30% (2000) 7.5% (2019) 4% (2009) 10% (2015) 15% (2015) 20% (2015) 22.5% (2021) 10% (2011) 24% (2013) 8% (2020) 16% (2020) ∼4.2% (2015) 15% (2020) 11% (2022) 10% (2015)

No Yes Yes Yes Yes Yes Yes Yes Yes No Yes No Yes Yes Yes Yes Yes Yes Yes No Yes Yes

Yes No Yes Yes Yes No No No Yes No Yes Yes Yes Yes No Yes Yes Yes Yes No Yes No

EBL

(distributed generation) (solar) (class I/II technologies) (vintage)

(class I/II technologies) (biomass, community wind) (community wind) (solar) (solar, class I/II technologies) (distributed generation) (solar) (vintage) (goal, non-wind) (solar, class I/II technologies)

Source: Wiser et al. (2007).

the RO target provides additional economic incentives. That explains why ROC prices were higher than the buy-out price in the ﬁrst years. This situation can be expected as long as the market is short of electricity from RES. Figure 12.12 depicts the number of ROCs issued in United Kingdom between October 2002 and March 2005 by technology and country.20 Clearly, in England the cheap options landﬁll gas and biomass coﬁring dominate. In Wales and Scotland onshore wind and hydro power are also among the preferred options. Italy introduced its TGC scheme in 2002, obligating all producers and importers of electricity to supply 2% of their power from new renewable electricity, with exceptions for CHP plants, renewables, and companies generating less that 100 GWh. Today the situation in Italy is similar to the British case. TGC prices are among the highest in Europe21 (Fig. 12.11) mainly because they expire in 8 years. Two types of TGC are on the market: those from qualiﬁed facilities and those sold by the market operator Gestore Servizi Elettrici (GSE), who trades the certiﬁcates issued to generators under contract with the previous FIT program at a price calculated each year. The role of the certiﬁcates sold by 20

The UK ROCs system has not directly affected small generators, e.g., roof PV, small wind, and small hydro power. 21 The high price of the TGCs sold by GSE is due to the mechanism for the price setting. When the generation of low-price sources like hydro power is low, the weighted average of the price paid for former feed-in contracts is higher and so the price of certiﬁcates. The same happens when the avoided cost, which has a cost factor related to the fuel prices, rises as happened in 2006.

Competitive Electricity Markets

Average level of TGC prices(/MWh)

448

120 100 80 60 40 20 0 2003

2004

2005

Italy

UK – RO

Flanders

Wallonia

Sweden

Japan

US – Massachusetts

US – Texas

Fig. 12.11. Comparison of TGCs price levels in selected countries (since TGC prices in the United States vary by state, Massachusetts and Texas are used as examples). Sources: Held et al. (2006), METI (2006), Wiser et al. (2007).

Rocs issued (1 ROC = 1MWh)

10 000 000 9 000 000 8 000 000 7 000 000

EBL

Scotland Wales England

6 000 000 5 000 000 4 000 000 3 000 000 2 000 000 1 000 000 0 Landfill Biomass Biomass Sewage Onshore Offshore Hydro power gas co firing (others) gas wind wind <20 MW

Technology

Micro hydro power

PV

Fig. 12.12. Number of ROCs issued in the period April 2002–March 2005 by technology and country (note: 1 ROC = 1 MWh). Source: OFGEM (2006).

GSE is to reconcile the previous FIT scheme with the new – a delicate task since the quota of the obligations has to be managed to avoid the supply of certiﬁcates from new plants from exceeding demand, which will result in zero price for the certiﬁcates. To facilitate this control and give elasticity to the supply, banking of certiﬁcates has been allowed for 2 years. In Belgium, two TGC schemes have been in existence in parallel since 2002, one in Flanders and the other in Walloon. The former was designed to promote energy generation from waste, biomass, and wind and it was clear from the beginning that the limited market may result in liquidity problems. Currently, TGC prices in Flanders are among the highest in Europe (Fig. 12.11). But as shown in Fig. 12.25, if the windfall proﬁts due to the promotion of prior capacity are taken into account, the additional costs

Promoting Electricity from Renewable Energy Sources

449

Cumulative renewables – nameplate capacity, MW

25 000

20 000

Capacity built in RPS states 15 000

10 000

5 000

Capacity Built in Non-RPS States

0 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004

Fig. 12.13. Cumulative non-hydro capacity from renewables in the United States. Source: Black and Veatch (2006, personal communication with Ryan Wiser).

EBL

for customers for generating new electricity from RES increase to about 18 cent/kWh (Verbruggen, 2005). The current penalty for not fulﬁlling the quota, of the order of 100–125 EUR/MWh, is not considered a major barrier since it is in the same range as the actual certiﬁcate prices (Fig. 12.11). In the case of Sweden, new RES capacity increased signiﬁcantly in 2004 and 2005 when certiﬁcate prices were low (Figs 12.13 and 12.26). The Swedish quota system allowed some old capacity to qualify for certiﬁcates22 resulting in a free-rider problem and generating windfall proﬁts for plants constructed before the TGC scheme went into effect. The availability of additional tax incentives and investment subsidies, especially for wind power plants, has contributed to the problems. 12.4.3.2. Renewable portfolio standards in the United States In the United States RPS have become the most common instrument for promotion of renewables at the state level, with 21 states and the District of Columbia adopting such schemes (Langniss and Wiser, 2003; Wiser et al., 2007). These schemes collectively encompass 40% of electricity supply in the United States and set minimum standards for renewable energy in the energy mix. As detailed in Wiser et al. (2005), the design of quotas varies considerably across states (see Table 12.5). The full effect of these RPS policies has not yet registered since only a few states have more than 5 years of experience and some of the quotas have been set but have not yet produced results. In the past few years, however, these policies have begun to have a 22

Recently, this system has been modiﬁed and currently mainly new capacities qualify for certiﬁcates traded.

450

Competitive Electricity Markets

sizable impact on the renewable electricity market in the United States.23 Figure 12.13, for example, shows that roughly half of the non-hydro renewable energy capacity additions since 2000 in the United States have occurred in states with quota obligations, most from wind power. Importantly, because of technology set-asides that exist in a number of states, a growing amount of solar energy is also being supported by these obligations. By some estimates, these quota systems could result in the installation of over 40 GW of new RES capacity by 2020 generating roughly 3% of projected US electric sales. In general, the most successful quota systems in the United States are those that have required or motivated long-term contracting with RES developers, with short-term contracts of unbundled TGCs used as a secondary compliance tool. These long-term contracts are typically the result of a competitive bidding procedure conducted by retail suppliers obligated to meet the quota. Though a signiﬁcant degree of contract failure has occurred in some states and government oversight has been required in others, many schemes appear to be functioning with efﬁciency and effectiveness. In some states where unbundled, short-term trade in TGCs dominate the market, however, problems similar to those in the United Kingdom have arisen, with TGC prices set by the penalty level, rather than based on market forces. In these instances, renewable electricity projects struggle to receive ﬁnancing, despite potentially high prices and profits due to the risk involved. This has especially been the case in restructured electricity markets,24 where load obligations are uncertain, and retail suppliers have typically been reluctant or unable to enter into longer-term contractual arrangements for electricity from renewables or unbundled TGCs.25 It remains to be seen whether this factor will complicate new project development in the long term and whether the resulting aggregate cost of the quota will be acceptable. Though short-term trade in TGC is not common in many of the RPS markets in the United States, several states have TGC markets that are sufﬁciently liquid to have transaction price data available (in the United States, TGCs are referred to as renewable energy certiﬁcates, or RECs). These states are typically those in which both retail electric competition and liquid wholesale electricity markets exist. Figure 12.14 presents monthly data on the average price of RECs in six different states and the District of Columbia. Clearly, TGC prices in the United States have experienced considerable variations among markets, and even within a single market over time. TGC price differences across markets reﬂect dramatically different state RPS designs as well as differences in available resources, vintage and geographic eligibility rules, the level of the RPS compliance target, the cost and availability of renewable generation in the region, and the level and design of any cost cap, to name a few. Variations in TGC prices within a given market and over time reﬂect the inﬂuence of changes in RPS rules or expectations of those rules, the actual and/or expected speed of renewable energy development relative to the RPS targets, and the degree of competition for renewable energy from other states or from the voluntary green power market, among other factors.

EBL

23

It is important to note that most of the state RPS policies work in combination with federal tax incentives available for RES projects. In addition, many of these policies are applied in still-regulated electricity markets in which regulated utilities solicit long-term renewable electricity contracts to comply with the standards. 24 Some states have restructured their electricity markets in the United States to allow competition among retail electricity suppliers, while others have not. 25 Customers in restructured markets can switch suppliers, adding to retailer uncertainties in securing long-term contracts.

Promoting Electricity from Renewable Energy Sources

$60

$300 Massachusetts (left axis) Texas (left axis) DC Class I (left axis) $250

$200

$30

$150

$20

$100

$10

$50

$0

$0

ct

O

Au

g-

02 -0 D 2 ec Fe 02 b0 Ap 3 r-0 Ju 3 nAu 03 g0 O 3 ct D 03 ec Fe 03 b0 Ap 4 r-0 Ju 4 nAu 04 g0 O 4 ct -0 D 4 ec Fe 04 b0 Ap 5 r-0 Ju 5 nAu 05 g0 O 5 ct -0 D 5 ec Fe 05 b0 Ap 6 rJu 06 nAu 06 g0 O 6 ct -0 6

$40

Average Monthly REC Prices ($/MWh)

CT Class I (left axis) NJ Class I (left axis) MD Class I (left axis) NJ Solar (right axis)

$50 Average monthly REC prices ($/MWh)

451

Fig. 12.14. REC prices in selected states. Source: Wiser et al. (2007).

EBL

In a number of states, a variety of design pitfalls including the following have been experienced, causing quota systems to under-perform: • • • • • •

uncertainty in the duration or design of the quota policy; quota targets and eligibility rules that do not require new renewable capacity development; unclear or inadequate enforcement of the quota; quota targets that are too aggressive to be achieved; extensive exemptions of potential retail suppliers obligated to meet the quota; and inadequate compliance ﬂexibility.

Overall, experience with the quota schemes in the United States has been decidedly mixed. Where the RPS scheme is mainly based on competitive bidding and long-term contracts with suppliers, they appear to be working effectively and efﬁciently. As previously mentioned, RPS programs appear particularly problematic in restructured electricity markets where retailers are uncertain of their future load obligations and are therefore sometimes unwilling to enter into long-term contracts. Despite the mixed experiences, quota policies are likely to remain the predominant form of support for renewables in the United States, at least in the near term. 12.4.3.3. Renewable energy schemes in Japan In 2003, Japan introduced an RPS scheme requiring that approximately 1.35% of each retail supplier’s sales in 201026 come from eligible RES, deﬁned as PV, wind, biomass, geothermal, and small hydro power (1 MW or less), rising to 1.63% by 2014. Electricity 26

All dates are ﬁscal year (from April to March) in Japan.

452

Competitive Electricity Markets

from PV is credited at two times the value from 2011 to 2014. To be certiﬁed the renewable electricity must be sold to the grid. The total target has been set to increase from 3.3 TWh in 2003 to 12.2 TWh in 2010 and 16 TWh by 2014. The targets are low27 compared to those in the Europe and the United States, partly because large hydro power and geothermal are ineligible under the scheme and also because a considerable amount of electricity generated from biomass is consumed for self-use. As with other RPS schemes, retail suppliers and renewable generators may trade certiﬁcates. Also, banking and borrowing of certiﬁcates up to 20% of the target are allowed. The maximum price of the certiﬁcate is set at 11 JPY/kWh (approx. 9 US cents/kWh). The total amount of RES supplied in 2005 was 5.6 TWh,28 which exceeded the actual target of 3.8 TWh. The targets from 2006 to 2009 were revised upward by 4 TWh in total as a part of the review process conducted in 2006. Since the enactment of the RPS scheme, renewable generation has steadily increased (Fig. 12.4), a trend that is expected to continue (Nishio and Asano, 2006), while prices have declined. The certiﬁcates were traded at a relatively stable price range of around 5 JPY/kWh (approx. 6 US cents/kWh) from 2003 to 2005 (Fig. 12.11), presumably because the transaction prices are determined by taking the banking into consideration from a long-term viewpoint. 12.4.4. Promotion of PV Several large-scale programs to promote PV have been implemented in different parts of the world, the cumulative effects of which are shown in Fig. 12.15. Three of the most signiﬁcant are brieﬂy described.

EBL

12.4.4.1. Japanese residential PV dissemination program A subsidy program for promotion of PV in the residential sector in Japan was launched in 1994. By the end of 2005 when the subsidy program was terminated29 it had resulted in over 1 GW of new capacity and the average cost of a residential PV system had dropped by more than half (Fig. 12.16). The subsidy program succeeded in creating a market, which accounted for less than 1% of the whole residential market, but was large enough to justify large investments in mass-production facilities. Several factors contributed to success of the program including a consistent technologypush policy30 (Kimura and Suzuki, 2006). However, it is not clear if the momentum of the program will continue following the termination of investment subsidies in 2005. It is not likely that additional promotion schemes in the residential sector will be launched by the government, currently focusing on R&D support for innovative PV technologies as well as subsidies in the non-residential sector. Moreover, the future role of the voluntary net metering schemes currently offered by the power companies is not certain. 27

Renewables currently account for roughly 10% of Japanese generation. The mix was PV 0.46 TWh, wind power 1.91 TWh, biomass 2.50 TWh, and small hydro power 0.70 TWh. 29 Several local governments are continuing ﬁnancial support programs. 30 The Japanese government has been providing larger and more stable R&D budgets for PV over the last quarter century than other major producing countries such as the United States and Germany according to IEA’s R&D statistics. Also, for reference, the overall budget related to new energy technologies by METI (Ministry of Economy, Trade and Industry) has tripled from 48 billion JPY in FY1996 to 156 billion JPY in FY2006. 28

Promoting Electricity from Renewable Energy Sources

Cummullative installed capacity (MWp)

California

Germany

453 Japan

2500

German EEG amendment

2000

1500

Californian CERP 1000

German HTDP

Japanese RPVDP

500

0 1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

Fig. 12.15. Cumulative installed PV capacity in selected markets. Source: Lopez and Haas (2007).

300 250

2000

EBL

1500

200 150

MWp

Thoudand JPY/kWp

2500

1000 100 500

50

0

0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 System cost

Subsidy

MWp/year

Fig. 12.16. The Japanese residential PV promotion program. Source: Lopez and Haas (2007), NEF (2006).

12.4.4.2. German rooftop PV programs The ﬁrst major promotional program for residential PV was the 1000 roofs program launched in Germany in 1989 and completed in 1994. The scheme resulted in installation of PV systems with an average size of 2.6 kW and a total capacity of 6.15 MW on some 2250 German roofs. Average system cost was US$15 000/kW with subsidies covering 70% of the investment costs. An expansion of the ﬁrst scheme, called the 100,000 Roofs Program, was launched in 1999 with the aim of reaching 100 000 installations with an average size of 3 kW for a total installed capacity of 300 MW. Low-interest loans were provided as the main inducement,

Competitive Electricity Markets

454 7000

400 000

6000

350 000

€/kWp

250 000 4000 200 000 3000

kWp/year

300 000

5000

150 000 2000

100 000

1000 0 1999

50 000

2000

2001

Subsidy

2002

System price

2003

0 2004

kWp/year

Fig. 12.17. The German 100,000 roofs program. Source: Lopez and Haas (2007).

initially set at 0% and with a payback time of 10 years. The initial response to the program between 1999 and 2002 was rather modest (Fig. 12.17) and the program suffered from a number of stops and starts. In 2000 the interest rate was raised to 1.8% and favorable FITs of 50.6Ecent/kWh were introduced. The combination of the low interest and FIT led to an impressive uptake resulting in deployment of 261 MW.

EBL

12.4.4.3. California’s PV programs California is the market leader for grid-connected PV in the United States driven by a mixture of state and local incentives as well as plentiful sunshine. Historically, PV technology has been supported through capital cost rebates – denominated in $ per watt – offered to PV system installers or owners to buy down the installed cost of solar installations, though more recently performance-based incentives have also been used. Incentive programs have been supported by the California Energy Commission (CEC) and the California Public Utilities Commission (CPUC), as well as the state’s publicly owned utilities. The CEC has administered a PV incentive program called the Emerging Renewables Program since 1998. As of the end of 2005, the CEC had paid out incentives to over 15 000 PV systems, totaling 62 MW in capacity. The CPUC’s Self-Generation Incentive Program (SGIP) began accepting applications in 2001 and offered rebates for customer-sited PV systems of at least 30 kW in size31 and installed by customers taking electric or gas service from one of the state’s private utilities. As of the end of 2005, the CPUC had paid out incentives to 403 PV systems, totaling 49 MW. Over time, both the CPUC and the CEC programs have altered the structure and size of their incentives for PV installations. The CPUC initiated its incentives at $4.5/W and dropped the incentive level to $3.5/W in December 2004; the CPUC further reduced the incentive to $2.8/W for applications received after December 2005. The CEC’s standard 31

Systems can exceed 1 MW in size, but the rebate only applies to the ﬁrst 1 MW.

Promoting Electricity from Renewable Energy Sources

455

incentive started at $3/W, increased during the state’s electricity crisis to $4.5/W, and then declined to $2.6/W. At the beginning of 2007, California’s solar programs were restructured, including a move towards performance ($/kWh) based incentives. As shown in Fig. 12.18 customer response to these incentives was disappointing at ﬁrst, with relatively little PV capacity being added through 2000. Demand for PV increased substantially after 2000, mainly as a result of the state’s electricity crisis in 2001 and in response to the higher rebates offered at that time (Bolinger and Wiser, 2002). Cost reductions under the program have been substantial, at least for smaller systems. Meanwhile, the state government in 2006 established the California Million Solar Roofs initiative, with a goal of encouraging 3000 MW of new solar PV systems through a longterm, sustained, declining incentive program. The CPUC – in conjunction with the CEC – has developed an implementation plan for this initiative, which includes performancebased payments for the majority of systems over 30 kW in size, replacing the current up-front rebates. The program is envisioned to be signiﬁcant is size (∼$3.2 billion) and stable (∼11 years) enough to signiﬁcantly reduce system costs over time.

12.4.5. Investment-based tax incentives A number of options have been used to promote renewables with ﬁscal instruments including • • •

lower VAT rate applied for renewable electricity systems; making dividends from RES investment exempt from income taxes; and tax credits for investments in RES.

EBL

These options have similar impact, acting as investment subsidies for new installations. Table 12.6 gives an overview of existing investment-based tax incentives in EU countries and the United States.

9000

20 000

8000

18 000 16 000

7000

€/kWp

12 000 5000 10 000 4000 8 000 3000

6 000

2000

4 000

1000

2 000

0 1999

2000

Subsidy

2001

2002

System price

2003

0 2004

kWp/year

Fig. 12.18. California’s Emerging Renewables Buydown Programme (CERBP). Source: CEC.

kWp/year

14 000

6000

456

Competitive Electricity Markets

Table 12.6. Investment-based tax incentives in various EU countries and the United States as of the end of 2006. Country

Investment-based tax incentives

Austria

Private investors get tax credits for investments in using renewable energies (personal income tax). The amount is generally limited to 2.929E per year 13.5–14% of RES investments deductible from company proﬁts, regressive depreciation of investments. Reduced VAT on building retroﬁt if energy efﬁciency is included (6% instead of 21%) The ﬁrst 3000 DKK of income from wind energy is tax-free Deduction of 15% investment costs with a maximum of 3000E per person. Reduced VAT (5.5%) on renewable equipment (not applicable to installation costs) Losses of investments can be deducted from the taxable income. This fact increases return on investments into wind projects Up to 75% of RES investments can be deducted Corporate tax incentive: tax relief capped at 50% of all capital expenditure for certain RES investments VAT reduced to 10% for investments in wind and solar; 36% deduction of PV, solar thermal and energy efﬁciency investments up to 54 000E (55% from 2007) Up to 30% of any type of investments on RES can be deducted with a maximum of 700 E per year. Reduced VAT (12%) on renewable equipment Corporation tax: 10% (up to 20% in some autonomous regions) tax liability instead of 35% for investments in environment-friendly ﬁxed assets EIA scheme: RES investors (most renewable energy systems) are eligible to reduce their taxable proﬁt with 55% of the invested sum Lower interest rates from Green Funds: RES investors (most renewable energy systems) can obtain lower interest rates (up to 1.5%) for their investments. Moreover dividends gained are free of income tax for private investors 30% federal investment tax credit available for solar installations (capped at $2000/system for residential users; up-capped for commercial systems) Favorable 5-year accelerated tax depreciation for most “new” RES-E A number of states offer their own income, sales, and property tax exemptions and incentives

Belgium

Denmark France Germany Greece Ireland Italy Portugal Spain The Netherlands

United States

EBL

Sources: Ragwitz et al. (2007), www.dsireusa.org.

One of the best-known tax incentives is the US federal production tax credit (PTC), which has been in place since 1994. The incentive is based on production, rather than investment, and is currently valued at 2.0 cents/kWh. Development of wind power in the United States in recent years has been strongly tied to the PTC combined with a number of state-level quota systems. Even where quota systems are not in place, wind development now occurs in some states based on the PTC alone. Unfortunately, the PTC has expired and been re-instituted with regularity making it difﬁcult for developers, investors, and ﬁnanciers to plan ahead, and resulting as a boom-and-bust cycle of wind development as shown in Fig. 12.19.

12.4.6. Mixed strategies: wind energy in Denmark In terms of large-scale integration of wind power in the electricity system, Denmark is in a class of its own. In 2005, nearly 20% of the country’s electricity consumption was

Promoting Electricity from Renewable Energy Sources

457

3000

Capacity (MW)

2500 2000 1500 1000 500 0 1997

1998

1999

2000

2001

2002

2003

2004

2005

Fig. 12.19. US wind power additions, by year. Source: www.awea.org/faq/instcap.html

produced with wind power.32 The western part of the Danish grid, which is not connected to the grid in the east, gets 24% of its electricity from wind power (www.ens.dk). The major reason for this impressive record is that wind power has had a prominent role in the Danish energy plans from 1990 and 1995. The target for wind in 2005 was an installed capacity of 1500 MW or around 10% of Danish electricity demand. This target was exceeded by a factor of two by 2003, where the installed wind capacity passed the 3000 MW mark (Meyer, 2004 a, b). Additionally, Denmark has enjoyed a stable legal framework and a favorable FIT scheme supported by successive governments. This created a stable investment climate in the 1990s and ensured that the overall energy policy did not change dramatically until a shift occurred in 2001 with the arrival of a liberal-conservative government. Another contributing factor was the introduction of a comprehensive wind atlas showing the local potential for wind energy in different parts of the country (Petersen et al., 1981). Prior to 1990s, the majority of renewable generators were cooperatives who enjoyed tax exemptions for their shareholders, guaranteed minimum price system and preferential treatment for the neighborhood. Starting in early 1990s Danish municipalities were forced to indicate sites suited for wind power generation. At that time many farmers saw an advantage in owning their own turbines as a ﬁnancial investment that could be written off on the business account of the farm. This possibility was not available for the cooperatives. As a result many of the new turbines in the late 1990s were owned by farmers and developers. Since 2001, anyone, including investors from abroad, may own wind turbines in Denmark. At the end of 2006 nearly 5500 wind turbines were operating in the country. In 2004, a political agreement was reached by the Danish Parliament to increase wind power capacity over the coming years by some 350 MW through a repowering scheme. Furthermore, the agreement included two tenders for offshore wind farms of 200 MW each, together with a decision to introduce full legal and ownership unbundling by separating transmission and production of electricity. This is expected to increase wind power’s share of Danish power generation to 25% by 2008. Beyond 2008 it is expected that most of the development will have to be offshore and by the replacement of older onshore turbines.

EBL

32

Adjusted to an average wind year.

458

Competitive Electricity Markets

The Danish Wind Associations has proposed a goal of 50% wind power by 2025 in the Danish energy mix with the installation of 200 MW per year. A recently published analysis from the association shows that wind power’s share of Danish electricity consumption could be increased from the 20% or 6.6 TWh in 2004 to 50% or 19 TWh while reducing the number of wind turbines by more than two thirds, from the current 5500 to 1750. The turbine types onshore are assumed to be 1, 1.5, and 3 MW machines, all commercially available today, while the offshore turbines are assumed to be 4 and 6 MW turbines.

12.5. What Works and Why? Reviewing the variety of schemes and instruments described above, one is tempted to ask whether these programs have been successful and if so by what measure? The two most important criteria are effectiveness and economic efﬁciency. Additional criteria include credibility of the scheme for investors and the reduction of generation costs over time. Table 12.7 provides a summary of the relevant performance parameters, which are further described below.

12.5.1. Effectiveness of policy instruments To examine the effectiveness, one must look at the relevant outcome, in this case the quantities generated or capacities installed and so on. To make relevant comparison among different countries, the ﬁgures must be examined by capita. Moreover, one must examine all new RES as well as speciﬁc types such as wind and PV. Figure 12.21 shows policy effectiveness of different policies for electricity generation from all new RES for the period 1998–2004 for the EU, United States, and Japan measured in terms of incremental amount of RES installed per year and capita. Not surprisingly, Denmark ranks the highest on this score with about twice as high renewable electricity deployed than the next ranked countries Finland, Sweden, Spain, and Germany. It should be noticed, however, that since 2003 the net increase in wind power capacity has been close to zero in Denmark. Many of the variations in Fig. 12.20 can be attributed to different promotion schemes such as the quota-based TGC system in Sweden as opposed to investment incentives in Finland and FITs in the other countries. Other factors also play a role such as the availability of inexpensive hydro power electricity in Nordic countries and plentiful supplies of cheap electricity from biomass. Moreover, progress was generally much slower in new EU member states than in the old EU-15 countries. Of the former, Hungary and Latvia showed the highest relative growth in the period considered. The United States and Japan deployed clearly less new RES electricity per capita than the EU-25. Looking at onshore wind (Fig. 12.21) the EU countries with the highest policy effectiveness during the considered period – Demark, Germany, and Spain – are the ones that applied ﬁxed FITs during the entire period 1998–2005, except for Denmark which had a change in 2001. The resulting high investment security as well as low administrative barriers stimulated a strong and continuous growth in wind energy during the last decade. By contrast, high administrative barriers in countries like France can signiﬁcantly hamper the development of wind energy even under a stable policy environment combined with reasonably high FITs. With respect to PV – currently one of the most expensive among renewable technologies – Germany and Japan show the highest effectiveness based on this particular measure

EBL

Table 12.7. Summary of performance parameters. Period of time analyzed

RES quantity deployed (W/cap year)

Magnitude of absolute support level

Decrease in support over time?

Risk for investors

Other important aspects

FIT and premium United States (PURPA) Denmark Germany Spain

1978–90

Medium

High

No

Low

1992–99 1998–2005 2002–05

High High High

No Yes Yes

Low Low Low

Austria

2002–05

High

Low Medium Low (ﬁxed option); medium (premium) Medium

No

Low

Portugal France

2002–05 2002–05

High Low

Low Medium

No No

Low Low

High administrative barriers

Low (quota not met) Low

High

Yes

Medium/high

Penalty too low

High

No

High

High (quota met) Low (quota not met) High (quota met) Medium (quota not met) Low (quota met)

Low

Constant

Medium

High

No

Medium/high

Low

No

Low/medium

Time of validity of RES plants for certiﬁcates too low (8 years) Windfall proﬁts due to some old capacities also qualifying for certiﬁcates Low penalty, windfall proﬁts due to some old capacities also qualifying for certiﬁcates Low with long-term contracts available

High

No

High

Few longer-term contracts available TGCs

Low

Low

RPS and quota-based TGC United Kingdom 2003–05 (RO) Italy 2003–05 Sweden

2003–05

Belgium

2003–05

Texas

2003–05

Massachusetts

2003–05

Japan

2003–05

Tendering United Kingdom (NFFO)

1990–98

Medium

EBL

No

Yes

Support level to high because of parallel investment subsidies

Low

Low after selection

Capacities to low

Competitive Electricity Markets

460

Fig. 12.20. Policy effectiveness of support measures for electricity from new RES (excluding hydro power) measured in additional kWh per year and per capita for the period 1998–2004 in the EU, United States, and Japan. Sources: EUROSTAT (2007), IEA (2006b), METI (2007), Black and Veatch (2006, personal communication with Ryan Wiser).

Feed-in tariff

Quota / TGC

Tender

UK

SI

SK

0 SE

0 PL

5

PT

5

NL

10

MT

10

LT

15

LU

15

IT

20

LA

20

IE

25

HU

25

GR

30

FI

30

FR

35

ES

35

EE

40

DK

40

CZ

45

DE

45

CY

50

AT

50

BE

Average additional wind capacity per year and capita between 1998 and 2005 (W/(year*capita)

EBL

EU-Avg United States

Japan

Tax incentives / Investment grants

Fig. 12.21. Policy effectiveness of onshore wind measured in additional capacity per year and per capita in the period 1998–2005 in the EU, United States, and Japan. Sources: EUROSTAT (2007), IEA (2006b), METI (2007), Black and Veatch (2006, personal communication with Ryan Wiser).

Average additional electricity generation from Solar Energy per year and capita between 1998 and 2005 (W/(year*capita)

Promoting Electricity from Renewable Energy Sources

461

6

6

5

5

4

4

3

3

2

2

1

1

0

0 AT BE CY CZ DE DK EE ES FI FR GR HU IE IT LA LT LU MT NL PL PT SE SI SK UK

Feed-in tariff

Quota / TGC

Tender

EU-25 United Japan States

Tax incentives / Investment grants

Fig. 12.22. Policy effectiveness of PV electricity support measured in additional capacity per year and per capita in the period 1998–2005 in the EU, United States, and Japan. Sources: EUROSTAT (2007), IEA (2006b), METI (2007), Black and Veatch (2006, personal communication with Ryan Wiser).

EBL

(Fig. 12.22). Obviously, generous FITs, as in Germany and Luxembourg, combined with net metering and rebates in Japan, produce results. 12.5.2. Economic efﬁciency In examining economic efﬁciency three parameters are of interest: absolute support levels, total costs to society, and dynamics of the technology. As an indicator in the following the support levels are speciﬁcally compared for wind power in the EU-15.33 Figure 12.23 shows that for many countries the support level and the generation costs are very close. Countries with rather high average generation costs frequently show a higher support level. A deviation from this trend can be found in the three quota systems in Belgium, Italy, and the United Kingdom, for which the support is presently signiﬁcantly higher than the generation costs. The reasons for the higher support level expressed by the current green certiﬁcate prices may differ. Main reasons are risk premiums, immature TGC markets, and short validity times for the certiﬁcates, which apply to Italy and Belgium. For Finland, the level of support for onshore wind is too low to initiate any steady growth in capacity. In the case of Spain and Germany, the support level indicated in Fig. 12.23 appears to be above the average level of generation costs. However, the potentials with rather low average generation costs have already been exploited in these countries due to the recent successful market growth. Therefore a level of support that is moderately higher than average costs seems to be reasonable even if it results in windfall proﬁts for some 33

A comparison of all new RES would provide too broad ranges for generation costs as well as for support measures.

Competitive Electricity Markets

462

160

160

Minimum to average generation costs (€/MWh) 140

Minimum to average generation costs (US $/MWh)

120

120

Average to maximum support level (US $/MWh)

100

100

80

80

60

60

40

40

20

20

Average to maximum support level (€/MWh)

140

0

0 AT BE CY CZ DE DK EE ES FI FR GR HU IE IT LA LT LU MT NL PL PT SE SI SK UK

United States

Japan

Fig. 12.23. Support levels for onshore wind (average to maximum) in the EU, United States, and Japan in 2005. Minimum to average generation costs are shown because this range typically contains presently realizable potentials which investors would normally deploy in order to generate electricity at minimum costs. Furthermore, the maximum generation costs can be very high in each country so that showing the upper cost range for the different RES-E would affect the readability of the graphs. Source: Adapted from Ragwitz et al. (2007).

EBL

34

wind power owners. In an assessment over time also the potential technology learning effects should be taken into account in the support scheme. 12.5.3. Quantities vs costs of support Next the relation between quantities deployed and the level of support is analyzed. It is often argued that the reason for higher capacities installed is a higher support level. And it is accepted that the resource endowments of RES vary from country to country. Paradoxically, countries with highest support levels – Belgium and Italy for example – are among those with the lowest speciﬁc deployment (Fig. 12.24). On the other hand, high FITs especially in Germany and Spain are often named as the main driver for investments especially in wind energy. However, the support level in these countries is not particularly high compared with other countries analyzed here. 12.5.4. What triggers investments? Analyzing the various promotional strategies from the point of view of investors may allow a better understanding of program costs and help in the design of more efﬁcient schemes (Finon and Perez, 2007; Langniss and Wiser, 2003; Meyer, 2007). 34 Under TGC all the technologies receive the marginal cost, i.e., there is a higher proﬁt for some low-cost technologies (wind power, biogas in Italy). Under high quantitative targets this can result in higher costs for the electricity consumers.

Promoting Electricity from Renewable Energy Sources

463

14 BE

Support level (c/kWh)

12 10

TGC systems

IT

FIT systems

8 UK 6

AT

DE 4

ES

SE

2 0 0

20

40

60

80

100

120

140

160

kWh/cap/year Fig. 12.24. Effectiveness vs costs of promotion programs for electricity from RES in selected countries (2002–04). Source: own investigations.

EBL

FIT schemes and tendering instruments may work well in combination to promote both mature and less mature technologies. In fact there are similarities of producer–buyer arrangements with respect to these systems. The risks to purchasers and generators are largely alleviated in cases of FITs and tendering. This is due to long-term contracts ensured by governments. One should note, however, that in the tendering system, transaction costs are much higher for the developer than in the FIT system, due to the cost of preparation for the bid in the tendering process. In the quota-based TGC systems, multiple risks for the investors may emerge resulting in strong preference for long-term arrangements. The strong bilateral interdependence between developers and obligated purchasers may lead to long-term contracts and to vertical integration. Recourse to spot transactions of green certiﬁcates has turned out to be only marginal in determining the certiﬁcate price. For small- and medium-sized suppliers with uncertain demand, there remains a tension between the risk associated with the uncertainty of future loads and the certiﬁcate obligation as well as the efﬁciency of managing the risks by long-term contracting – as shown in the RPS programs which are set in some of the most liberalized US electricity markets. However, some authors (e.g., Lemming, 2003) argue that with regard to ﬁnancial risks, the TGC/quota systems may give incentives to renewable electricity developers to avoid contracts with forward ﬁxed price because the spot market will give higher prospects for proﬁt. Four elements play against this view. First, volatility and price risk are high because the size of the certiﬁcates market is small. Second, in the case where RES producers sell green electricity as two products (electricity sale on one hand and green certiﬁcates on the other), the risk to the green certiﬁcate price is added to the risk of the wholesale electricity price. Certiﬁcates banking which is supposed to help the obligated suppliers to respect

464

Competitive Electricity Markets

their quotas can increase the lack of liquidity in a period of tight supply of certiﬁcates. Third, transactional complexity which results from intermittence of RES generation also inﬂuences the choice of long-term contract. The absence of a purchase obligation on physical electricity (it is only a quota of certiﬁcates) reinforces the producer’s incentive to conclude long-term contracts in order to simplify transactions.35 Fourth, the price of certiﬁcates is affected by a number of risks, in particular the regulatory risk arising from an eventual alteration in the renewables portfolio of eligible technologies (adding a cheap technology – e.g., co-ﬁring or burning waste – may lower the prices because it increases the quantity of available certiﬁcates). In addition comes the risk of large actors exercising market power. So the RES producers have good reasons for negotiating contracts with buyers who are subject to quotas. Furthermore, it is important to underline that most of the quantity-based instruments (European TGCs, American RPS programs) win in effectiveness when they beneﬁt from reinforcement by subsidy on investment or on production In the United Kingdom, technology-speciﬁc investment grants selected by tendering for projects based on secondranked technologies complement TGC systems in order to mitigate their drawback in fostering variety in technological deployment. In the United States the quite recent combination of the renewed federal support by tax credit on production and RPS programs have had important revival effect as shown by the recent wind power capacity growth (see Fig. 12.19). The above reﬂections indicate that a long-term and stable policy environment for potential investors – with favorable economic support schemes – may be the key criteria for the success of developing renewables markets. 12.5.5. Cost evolution of technologies

EBL

The cost reduction of the renewable technologies is another important criterion for evaluating the efﬁciency of policy instruments in relation to technological learning. In the following the development of costs for onshore wind and PV are examined. The development of investment costs for wind turbines from the early 1980s until 2005 in Europe and the United States is shown in Fig. 12.25. In both regions, costs have dropped signiﬁcantly. In Europe, costs decreased from around 2500 EUR/kW in 1982 to 1500 EUR/kW in 1990 and further to below 1000 E in 2000. Since then costs have stagnated due to shortage of turbines in a fast-growing market. A similar trend can be observed in the United States where installed wind projects costs are shown extending back to 1982, including both proposed and online projects. A signiﬁcant drop in investment costs36 up until about 1996 can be seen, and followed by stagnation and, more recently a rise. Similar trends can be observed for PV system costs, where high global demand has resulted in high proﬁts for developers. The cost development for electricity generation from (small) PV systems in different countries is shown in Fig. 12.26 where the decreasing cost trend has been replaced by stagnation and even a slight increase in costs during recent years. This is mainly due to shortage of basic silicon material for the production of traditional PVs in a fast-growing market. Also, the recent stagnation of cost reduction implies that a further technology 35

Long-term contracts also deﬁne the party which pays the balancing costs: generally the obligated suppliers assume them. 36 The data in this graphic include information on 265 wind projects totaling 17 420 MW.

Promoting Electricity from Renewable Energy Sources

465

4500 4000 3500

Investment costs

3000

United States ($/kWp) 2500 2000 1500

Europe (EUR/kWp) 1000 500 0 1980

1985

1990

1995

2000

2005

2010

EBL

Fig. 12.25. Development of the investment costs of wind turbines in Europe (based on Denmark) and the United States. Data are in real 2006 prices. Sources: Nej et al. (2003), Morthorst (2006), Wiser and Bolinger (2007).

14 000

System costs (€/kWp)

12 000 10 000 8000 6000 4000 CA

Germany

Japan

2000 0 1994

1996

1998

2000

2002

2004

Fig. 12.26. Cost reductions of PV systems over time. Data are in real 2000 prices. Source: Lopez, Haas and Suna (2008).

466

Competitive Electricity Markets

innovation is required for PVs in order that they become economically competitive with conventional power plants.37 With respect to the price development in California Wiser et al. (2006a) argue that the lack of a sustained long-term policy commitment may be reducing the incentive for cost reductions. Moreover, PV prices have been ﬂuctuating alongside rebate levels: PV prices are increasing with rises in rebate levels – and falling with rebate drops. This means that subsidy variations, if not designed with care, may be used by retailers to increase their own proﬁts to the disadvantage of consumers in a sellers’ market. Perfect competition only exists in theory, and this should be taken into account when designing support schemes. 12.6. Conclusions Clearly, a wide range of policy instruments have been tried and are in place in different parts of the world to promote renewable energy technologies. The design and performance of these schemes varies from place to place, requiring further research to determine their effectiveness in delivering the desired results. The main conclusions that can be drawn from the present analysis are as follows: •

Generally speaking, promotional schemes that are properly designed within a stable framework and offer long-term investment continuity produce better results. Credibility and continuity reduce risks thus leading to lower proﬁt requirements by investors. • Despite their signiﬁcant growth in absolute terms in a number of key markets, the nearterm prognosis for renewables is one of modest success if measured in terms of the percentage of the total energy provided by renewables on a world-wide basis. This is a signiﬁcant challenge, suggesting that renewables have to grow at an even faster pace if we expect them to contribute on a signiﬁcant scale to the world’s energy mix.

EBL

Acknowledgments Some work presented in this chapter was ﬁnanced within the scope of projects of the European Commission (GREEN-X, OPTRES). However, the views expressed here are the sole responsibility of the authors and not of these institutions. The authors are grateful to the following persons for their valuable contributions and discussions: Assun Lopez-Polo, Demet Suna, Thomas Faber, Claus Huber, Gustav Resch, Mario Ragwitz. References Bolinger, M. and Wiser, R. (2002). Customer-Sited PV: A Survey of Clean Energy Fund Support. LBNL49668. Berkeley, CA: Lawrence Berkeley National Laboratory. European Commission [EC]. (2001). Directive 2001/77/EC of the European Parliament and of the Council of 27 September 2001 on the promotion of electricity produced from renewable energy sources in the internal electricity market. European Commission [EC]. (2005). The support of electricity from renewable energy sources COM (2005) 627 ﬁnal.

37 For a more detailed analysis of the development of the costs of different components see Lopez, Haas and Suna (2008).

Promoting Electricity from Renewable Energy Sources

467

Eurostat. (2007). Energy: Yearly Statistics – Data 2005. Finon, D. and Perez, Y. (2007). The social efﬁciency of instruments of promotion of renewables in the electricity industry: a transaction cost perspective. Ecol. Econ., 62, 77–92. Flavin, Ch. and Lenssen, N. (1994). Power Surge. New York: W. W. Norton. Haas, R. (2003). Survey on and review of promotion strategies for RES in Europe. ENER Forum 3. Haas, R., Eichhammer, W., Huber, C., et al. (2004). How to promote renewable energy systems successfully and effectively. Energy Policy, 32(6), 833–9. Haas, R., Meyer, N.I., Held, A., et al. (2007). A review of promotion strategies for electricity from renewable energy sources (mimeo, paper submitted). Held, A., Haas, R., and Ragwitz, M. (2006). On the success of policy strategies for the promotion of electricity from renewable energy sources in the EU. Energy Environ, 17(6), 849–68. International Energy Agency. (2006a). World Energy Outlook 2006, Paris. International Energy Agency. (2006b). Renewables Information 2006. With 2005 data. Paris: OECD/IEA. Kimura, O. and Suzuki, T. (2006). 30 years of solar energy development in Japan: co-evolution process of technology, policies, and the market. Paper prepared for the 2006 Berlin Conference on the Human Dimensions of Global Environmental Change: “Resource Policies: Effectiveness, Efﬁciency, and Equity,” 17–18 November 2006, Berlin. Komor, P. (2003). Renewable Energy Policy. New York: Diebold Institute for Public Policy Studies. Langniss, O. and Wiser, R. (2003). The renewables portfolio standard in Texas: an early assessment. Energy Policy, 31, 527–35. Lemming, J. (2003). Financial risks for green electricity investors and producers in a tradable green certiﬁcate market. Energy Policy, 31, 21–32. Lopez, A., Haas, R., and Suna, D. (2008). An international comparison of market drivers for gridconnected PV systems, Progress in PV forthcoming. Martinot, E., Wiser, R., and Hamrin, J. (2005). Renewable energy policies and markets in the United States. Prepared for the Energy Foundation’s China Sustainable Energy Program. San Francisco, CA: Center for Resource Solutions. Menanteau, P., Finon, D., and Lamy, M.L. (2003). Prices versus quantities: environmental policies for promoting the development of renewable energy. Energy Policy, 31(8), 799–812. METI. (2006). Status of the implementation of the RPS in FY2005 (in Japanese only). METI. (2007). RPS web site at www.rps.go.jp/ (in Japanese only). Meyer, N.I. (2003). European schemes for promoting renewables in liberalised markets. Energy Policy, 31(7), 665–76. Meyer, N.I. (2004a). Renewable energy policy in Denmark. Energy Sustain. Dev., VIII (1), 25–35. Meyer, N.I. (2004b). Development of Danish wind power market. Energy Environ., 15 (4), 657–74. Meyer, N.I. (2007). Learning from Wind Energy Policy in the EU: Lessons from Denmark, Sweden and Spain. European Environment, 17, 347–362. Mitchell, C., Bauknecht, D., and Connor, P.M., et al. (2005). Effectiveness through risk reduction: a comparison of the renewable obligation in England and Wales and the feed-in system in Germany. Energy Policy, 34(3), 297–305. NEF. (2006). Solar NEF web site at www.solar.nef.or.jp/ (in Japanese). Nej et al., (2003). EXTOOL – Publishable database, Lund University, Sweden. Nishio, K. and Asano, H. (2006). Supply amount and marginal price of renewable electricity under the renewables portfolio standard in Japan. Energy Policy, 34(15), 2373–87. Oosterhuis, F. (2001). Energy subsidies in the European Union. Institute for Environmental studies. Petersen, E.L., Troen, I., Frandsen, S., and Hedegaard, K. (1981). Wind Atlas for Denmark. Roskilde, Denmark: RISÖ National Laboratory. Ragwitz, M., Held, A., Resch, G., et al. (2007). OPTRES – Assessment and optimisation of renewable energy support schemes in the European electricity market. Stultgart, Germany: Fraunhofer IRB Verlag. Scheer, H. (2001). The Solar Manifesto. James&James, London. The Ofﬁce of Gas and Electricity Markets [OFGEM] (2006). Renewables Obligation: Third annual report. Ref: 35/06. Van der Linden, N.H., Uyterlinde, M.A., Vrolijk, L., et al. (2005). Review of International Experience with Renewable Energy Obligation Support Mechanisms. Petten: Netherlands. ECN-C-05-025.

EBL

468

Competitive Electricity Markets

Verbruggen, A. (2005). Flanders’ Tradable Green Certiﬁcates System Performance January 2002–May 2005. Belgium: University of Antwerp. Unpublished. Paper presented at the REALISE workshop in Milano 2005, Italy. Voogt, M.H., Uyterlinde, M.A., de Noord, M., et al. (2001). REBUS: Renewable Energy Burden Sharing (Main Report). Netherlands. ECN-C-01-03. Wiser, R. and Bolinger, M. (2007). Annual Report on U.S. Wind Power Installation, Cost, and Performance Trends: 2006. Washington, DC: U.S. Department of Energy. Wiser R., Porter K., and Grace R. (2005). Evaluating experience with renewable portfolio standards in the United States. Mitigat. Adaptat. Strateg. Global Change, 10, 237–63. Wiser, R., Bolinger, M., Cappers, P., and Margolis, R. (2006a). Letting the Sun Shine on Solar Costs: An Empirical Investigation of Photovoltaic Cost Trends in California. LBNL-59282. Berkeley, CA: Lawrence Berkeley National Laboratory. Wiser, R., O’Connell, R., Bolinger, M., et al. (2006b). Building a “margin of safety” into renewable energy procurements: a review of experience with contract failure (CEC-300-2006-004). California Energy Commission. Wiser, R., Namovicz Ch., Gielecki M., and Smith R. (2007). Renewables Portfolio Standards: A Factual Introduction to Experience from the United States. LBNL-62569. Berkeley, CA: Lawrence Berkeley National Laboratory.

EBL

Chapter 13 Distributed Generation and the Regulation of Electricity Networks DIERK BAUKNECHT1 AND GERT BRUNEKREEFT2 1

Oeko-Institut, Germany; 2 Jacobs University Bremen, Germany

Summary Distributed generation (DG) enjoys high policy priority in many countries and is likely to have a bright future. Yet, although DG can have many advantages, it does not always easily ﬁt into today’s centralized power systems. This chapter looks at electricity networks and the role of network regulation for the integration of DG. It is often argued that DG can reduce network costs. However, in many cases, networks need to adapt to DG, incurring additional costs. Even if the overall beneﬁts of DG are positive, additional network costs represent a disincentive for network operators to connect DG. The chapter discusses various regulatory options to overcome this disincentive and improve coordination between network and plant operator.

EBL

13.1. Introduction Traditionally, the electricity supply industry has been dominated by centralized power generation. The network infrastructure and its regulation have developed accordingly, with the distribution network being a mere passive distributor of power. In recent years, this paradigm has been challenged by smaller, decentralized power plants connected to the distribution system (distributed generation or DG), including combined heat and power (CHP) and renewables. Their advantages can include economic, environmental, and security-of-supply aspects. Many countries have targets and policies to promote DG. It is a likely scenario that the share of these plants will increase over time (e.g., IZT et al., 2004). While small amounts of DG can be integrated into an existing network with relative ease, the same cannot be said when DG as a percentage of total generation grows to a signiﬁcant level. DG can have beneﬁts for the environment, for individual DG operators, or for the development of competition, but represents a new challenge for a hitherto centralized electricity system. With the share of DG increasing, their integration into markets and networks becomes an important issue. 469

Competitive Electricity Markets

470

While the technical debate on DG integration is well underway, the question as to how the governance structure of the electricity system can best promote and integrate DG has received less attention. Rather than just designing support mechanisms for niche markets, network regulation and general market design should provide for an efﬁcient integration of DG. This chapter focuses on the effects of DG on the electricity network. While it is often argued that electricity generation close to consumers can reduce electricity network costs, this requires that DG plants are integrated into network operation and there will be many cases where the network will need to adapt to DG, incurring additional costs. Even in cases where the overall beneﬁts of DG are positive, additional network costs clearly represent a disincentive for network operators to connect DG plants to their grid. This chapter examines the role of distribution network regulation for the integration of DG. Distribution network operators (DNOs) and the regulatory framework in which they operate play a central role in integrating and facilitating DG plants. This refers both to the incentives of DNOs to connect DG plants and to the coordination problem between DNO and DG. The regulatory framework essentially rests on two pillars: ﬁrst, economic regulation of revenues or proﬁts and, second, the approach toward vertical unbundling of the networks and the competitive businesses. Both have a potentially strong impact on the efﬁcient integration of DG. This chapter is structured as follows. Section 13.2 sets the context and provides an overview of DG, its deﬁnition, status, and drivers. It also explains why DG is not just a generation issue, but will affect the electricity network and is therefore something to be dealt with by network regulation. Section 13.3 discusses how the network effects of DG inﬂuence the DNO under different regulatory regimes and provides an overview of coordination problems between DNO and DG plants. Section 13.4 presents a number of approaches to incentivize the DNOs to connect DG to its network. Section 13.5 provides conclusions.

EBL

13.2. Background on DG, Electricity Networks, and Network Regulation 13.2.1. Distributed generation: deﬁnition and status There are a number of different deﬁnitions for DG, referring to • • • • • • •

the the the the the the the

location in the network, plant capacity, power delivery area, technology, environmental impact, mode of operation, or ownership.

Except for the ﬁrst, none of these deﬁnitions can adequately capture the range of plants that can be subsumed under the heading of DG, nor do they provide a satisfying description of their common characteristic. What makes a plant “distributed” is neither its technology nor its environmental impact, but its location within the network. While power plants have traditionally been connected to the transmission grid, distributed plants are connected to either the distribution grid or on the customer side of the meter. The deﬁnition of the distribution grid, i.e., the voltage level it comprises, depends on the system at hand.

Distributed Generation and the Regulation of Electricity Networks

471

In many cases, DG will be quite small, below 1 MW or even below 1 kW. Assuming that the maximum voltage level operated as part of the distribution system is 110 kV, the maximum capacity for DG can be in the range of 100–150 MW. In this chapter, the deﬁnition of DG by Ofgem (2002) is adopted, namely “DG, sometimes called embedded generation, is electricity generation, which is connected to the distribution network rather than high voltage transmission network. It is typically smaller generation such as renewable generation, including small hydro, wind and solar power and smaller Combined Heat and Power.” Although DG plays an increasingly prominent role in political debates and future scenarios of the electricity system, their current contribution is limited in terms of total capacity and generation in most electricity systems. What is more, in most countries, that are leaving the “grace period” inherited from the monopolistic era with signiﬁcant overcapacity and entering an investment phase, investment is mainly in centralized plants. Nevertheless, the share of DG is growing, though slowly. According to The World Alliance for Decentralized Energy (WADE, 2005), the world market share increased from 7% in 2002 to 7.2% in 2004. In absolute terms, in this 2-year period, around 32.2 GW of DG was added and global installed capacity stood at 281.9 GW. Most of these plants are industrial CHP or district heating plants. In 2005, DG contributed about 10% to total power generation worldwide. In terms of the share of DG in total new generation, this has risen from 13.0% in 2002 to around 24.5% in 2005 (WADE, 2006). Figure 13.1 gives an overview on the share of DG in 2005 in a number of countries including Denmark where they represent more than 50% of generation. As Jenkins et al. (2000, p. 7) have pointed out, although DG provides only a fraction of overall peak demand in most systems, what is also relevant in terms of system operation

EBL

DE share as % of total power generation

60.0

50.0

40.0

30.0

20.0

10.0

Denmark Finland Netherlands Latvia Czech Hungary Germany Turkey Slovakia Poland Japan Portugal Austria India Canada SouthAfrica Estonia World Chile China Korea Lihuania Mexico Uruguay Luxembourg Greece Spain Belgium Italy UK Sweden Slovenia Indonesia Uganda Australia France US Thailand Brazil Ireland Argentina

0.0

Fig. 13.1. Proportion of total power generation from decentralized capacity. Source: WADE (2006).

Competitive Electricity Markets

472

Development from the 1980s to the 1990s Primary generation

Local generation

Central power plant DCHP unit Wind turbine

Fig. 13.2. The growth of distributed generation in Denmark. Source: Eltra (now energinet.dk).

is the generation at times of minimum system load, as well as the density of DG in parts of the distribution network. It may well happen that individual distribution systems with DG become net exporters of power during periods of low load if local generation exceeds local demand. Denmark is the country with the largest penetration of DG capacity (Figs 13.1 and 13.2). Both wind and CHP plants have a major share in generation capacity. More than 50% of the total production capacity is dispersed throughout local distribution grids of 60 kV voltages and below. Several distribution companies have an installed DG capacity several times higher than their total load and therefore need to export power (Lund et al., 2006). This development is mainly due to political support for renewables and CHP that has provided a stable environment for DG since the 1970s (Haas et al., this volume; van der Vleuten and Raven, 2006).

EBL

13.2.2. Distributed generation: cost and beneﬁts There are different factors that have led to an increasing interest in DG (Alanne and Saari, 2006; DoE, 2007; Pepermans et al., 2005; Swisher, 2002). The potential advantages of DG include environmental, security-of-supply, and economic aspects. Although electricity market liberalization can, in principle, be said to have opened the door for DG, in that it has lifted restrictions on building power plants and opened the generation market to new entrants, the drivers for an increase in DG – and an increase in expectations as to the future of DG – have so far been mainly technical progress of DG technologies and the creation of niches outside the liberalized market in which these new technologies can operate supported by government policy (e.g., priority dispatch and feed-in mechanism for renewables). DG has only rarely been recognized as a separate policy objective, and in most countries, there is no explicit DG policy. Policies on DG tend to be buried under broader policies to

Distributed Generation and the Regulation of Electricity Networks

473

promote renewables and CHP (see Haas et al., this volume). Renewables are often small scale and connected to the distribution network.1 The same holds for many CHP plants. The increasing share of DG thus to a large extent results from political efforts to promote renewables and CHP, mainly for environmental reasons. As for security of supply, DG will lead to fuel diversiﬁcation, increased reliance on indigenous sources (e.g., wind), and therefore a reduction of import dependency on primary fuels. The geopolitical perspective is a strong energy policy driver in many countries. The economic beneﬁts of DG are ambiguous and can be positive or negative depending on speciﬁc circumstances. The economic impact of DG can be separated into two broad categories: • •

network-related beneﬁts (infrastructure), energy-related beneﬁts (commodity).

Within each category, there can be a range of different beneﬁts to the plant operator, the DNOs, the transmission system operators (TSOs), the customers, and society at large. DG beneﬁts tend to be highly technology, site, and time speciﬁc. Energy-related beneﬁts can include the following: • •

Reliability beneﬁts through on-site backup power. Option values and avoidance of overcapacity: As with grid investments, investments in centralized plants are much lumpier than DG investments and are, therefore, more difﬁcult to gear toward actual demand developments. In traditional power systems, an increasing electricity demand was usually met by installing a new largescale power plant. In today’s liberalized markets, building power plants has become a signiﬁcantly riskier investment and there are more incentives to avoid overcapacities. Small-scale DG plants are better suited to respond to demand changes.2 As a result, they have a higher option value than large-scale plants (Swisher, 2001). • Sites for large centralized power plants are becoming increasingly scarce. Small-scale DG is more ﬂexible and reduces the siting problem.

EBL

For the analysis in this chapter, the network-related costs and beneﬁts of DG are most relevant. The deﬁnition of DG as generation connected to the distribution system or on the customer side of the meter already indicates that DG need to be analyzed “in connection” with the network. Distributed plants are not just smaller plants but plants connected to a network level that was not designed for that purpose. In current electricity systems, the distribution system is different from the transmission system in a number of dimensions (Cardell and Tabors, 1997)3 : •

Distribution systems were designed to transport power from the transmission system to the consumers. They do not normally contain active sources of supply, and their protection systems may need to be adapted to accommodate generators. • Transmission systems are highly meshed, whereas distribution systems are characterized by a radial or a looped grid architecture. As a consequence, there are normally only one or maybe two paths to each bus, as compared to several alternative paths 1

The main exceptions are offshore wind parks. See Zarnikau, this volume, on the role of demand participation in liberalized electricity markets. 3 For a detailed description of the technical issues of DG network connections, see Jenkins et al. (2000). 2

474

Competitive Electricity Markets

in the meshed transmission system. Power ﬂows are unidirectional from higher to lower voltage levels. • The high-voltage lines of transmission systems have a low resistance as compared to the lower voltage lines in (rural) distribution system. As a consequence, if a generator gets connected to the low-voltage distribution system, this can signiﬁcantly affect the voltage levels. • In distribution systems, and especially the low-voltage ends, there are usually no control systems available that are standard on the transmission level. As a result, it is difﬁcult for DNOs to inﬂuence the operation of generators and maintain system stability. In a liberalized market with unbundled network operators and independent generators, this becomes even more of a problem, as generators are mainly concerned with responding to market signals or the maximization of revenues from support mechanisms rather than network requirements. As a result of these differences between the transmission and the distribution networks, connecting plants to the distribution level affects the system in a way that was not known before and requires DNOs, who are usually neither used nor equipped to handle large amounts of generation on their network, to deal with new problems. Given that distribution networks have generally not been designed to accommodate DG, it is not surprising that connecting DG to this network level can entail additional costs. First, there are costs for connecting the new plant to the grid, depending on the location of the DG. Choosing a DG location close to an existing grid may reduce connection costs. In addition to the connection costs, there are costs for reinforcing the grid beyond the point of connection to accommodate the additional load, e.g., if DG generation exceeds local demand and additional grid capacity is required to export electricity to the transmission grid. This is, for example, the case in Germany, where the high increase and regional concentration of wind capacity has led to a demand for additional network capacity (Burges and Twele, 2005). The same has happened in Denmark, where the costs of network reinforcement triggered by distributed wind and CHP plants between 1992 and 2001 amount to DKK 630 million (ca. US$ 76 million, 2001 exchange rate). This corresponds to an average of DKK 300 000 (ca. US$ 36 000) per MW for wind power and DKK 500 000 (ca. US$ 60 000) per MW for CHP. In comparison, the cost of building an onshore wind turbine was ca. DKK 6–7 million per MW (ca. US$ 0.72–0.84 million) in 2003 (Bach et al., 2003). What should not be neglected are the additional transaction costs that DG entails for the DNO. With DG, the DNO needs to deal with additional counter-parties, and given the relatively small size and distributed ownership of DG, this makes the DNO’s business more complex. These costs tend to be all the higher, the more DG is to be integrated into network planning and operation, as compared to a “ﬁt-and-forget” approach. While DNOs often consider DG as a new problem they have to deal with, there can also be beneﬁts to the system. DG plants can provide beneﬁts such as distribution capacity cost deferral, reliability improvements, or operational cost savings, resulting, for example, from reduced losses. DG is also likely to reduce the (peak) load a DNO needs to buy from the transmission system. In the United States, a number of studies on the costs and beneﬁts of DG have been conducted. A seminal work comes from Little (1999) and Gumerman et al. (2003). Iannucci et al. (2003) review 31 studies of the effects of DG and note that deferral of distribution network investment is the most important issue in 24 of 31 cases, followed by transmission network deferral and energy savings. These studies stress the beneﬁts of deferred investment in both transmission and distribution networks, as DG can be a substitute for

EBL

Distributed Generation and the Regulation of Electricity Networks

475

0.24

0.20

0.16 Costs to meet new 0.12 demand ($/kWh) 0.08

0.04

0

Generation, Transmission Generation No constraints transmission, and distribution constrained and distribution constrained constrained Central plant

All scenarios

Distributed generation

Fig. 13.3. Costs and beneﬁts of DG. Source : Little (1999), Fig. 3.1.

EBL

new lines (see also DoE, 2007: section 3.6). These effects can be substantial. For example, Little (1999) has estimated typical grid-side beneﬁts at US$ 55/kW per year for avoided increases in system capacity and US$ 30/kW per year for the deferral of transmission and distribution upgrades. Taken together, this translates into US$ 16/MWh, given a 60% load factor. The typical cost advantage of DG is depicted in Fig. 13.3, which compares the cost of installing DG (including secondary distribution investment) with the cost of central plant. The cost of the central plant scenarios in turn depends on the required power plant expansion and network upgrade. If generation and network capacity are currently sufﬁcient, then obviously installing DG has no cost beneﬁts. If, however, new DG avoids new central power plant or network expansion, the cost of DG may be less than the cost of the centralized option. The ﬁgure suggests that deferred network investment is critical. DG loses the comparison of the costs of generation only but starts to win the comparison if deferred network investment gains relevance. By and large, these results are conﬁrmed by the other studies mentioned above. The studies mentioned above are typical for the United States. Switching to a European context, DG may play a role in deferring transmission expansion, but constraints in the distribution network are rather unlikely. In most areas, the distribution network is large and well developed. What could be a driver for DG, though, is that in many cases in Europe, distribution network tends to be rather old and needs to be replaced in due time. However, it needs further investigation whether DG could signiﬁcantly substitute network renewal. Again, this is likely to be very case speciﬁc. For example, Cao et al. (2006) have modeled the Finish and the UK distribution systems and have come to the conclusion that “comparing the values of incremental network investment costs with incremental

476

Competitive Electricity Markets

potential beneﬁts of DG on network capacity replacement, we can see the investment saving due to the network reinforcement deferral can not offset the network upgrading cost for each scenario. Generally higher beneﬁts occur in relatively low DG penetration levels with lower density scenarios.” To sum up, the network impact of DG can be either positive or negative. The empirical effects are extremely case sensitive. How DG affects the DNO in practice depends on the details of each DG installation and the network. Even in those cases where there are network costs, they may be outweighed by non-network-related beneﬁts. And if it is a political objective to increase the share of DG, e.g., for environmental or security-of-supply reasons, this should be done at least cost. This needs to be taken into account when considering the design of network regulation.

13.2.3. DG and the development of electricity networks The focus of this chapter 13 is on connecting DG to the existing network. However, the more DG capacity increases, the more the network needs to adapt to be able to accommodate these generators. While the electricity network has been more or less stable over many decades, an increasing DG penetration will impose signiﬁcant challenges, both in network operation and in development. There is the assumption that the grid needs to be transformed if electricity generation is to become distributed and more sustainable. This is, for example, expressed in the Sussex Energy Group’s response to the latest UK energy review (Sussex Energy Group, SPRU, 2006): “The electricity network will need fundamental restructuring if the need to rely on more distributed sources of power is taken seriously. Much of this infrastructure was put in place 30–50 years ago and there is therefore a real opportunity to introduce a replacement system capable of responding to low carbon challenges.” Exactly how the future grid will look like that is capable of accommodating an increasing share of DG is unclear. There are a number of concepts and scenarios that go beyond removing barriers for individual plants but aim at changing the network philosophy to accommodate an increasing share of DG, e.g., active networks and microgrids (Coll-Mayor et al., 2006; EA Technology Ltd, 2001; European Commission, 2003; Strbac et al., 2007; Varming et al., 2002; see also http://www.epri-intelligrid.com). Some of these concepts represent long-term visions, while others are already being implemented, e.g., in Denmark. Table 13.1 provides an overview. For example, van Overbeeke and Roberts (2002) have proposed the “active networks” concept. In this approach, the current passive distribution networks that simply transport electricity from the transmission grid to the ﬁnal customers will be replaced by actively managed networks. An important feature of active networks is that they interact with their customers, i.e., DG plants can be controlled to adapt to the network situation. They argue that this is both technically and economically the best way to facilitate DG in a deregulated electricity market. Cao et al. (2006) have shown in a model-based study that active management of the distribution network enables a higher share of DG in a given network, and under certain DG scenarios (depending on density, penetration, and whether it is a rural or urban network), this can be more cost-effective than upgrading the network capacity. The International Energy Agency (IEA, 2002) describes the evolution of a decentralized system by distinguishing between the following three stages, where the share of DG increases from stages 1 to 3. The IEA characterizes the three stages as follows:

EBL

Table 13.1. Overview on innovative network concepts.

Objective

How?

Active distribution networks

Virtual power plants

Power cells (Denmark)

Microgrids

Connect more DG to existing network assets Increase utilization of existing networks Local and coordinated control of voltage, ﬂows, and fault levels Make distribution more intelligent, instead of investing in network primary plants (wires)

DG to trade on markets and provide network control and system support Aggregation of DG plants, increasing diversity and predictability

Integrate DG into system operation to achieve a higher level of DG within the existing infrastructure

Utilize DG to reduce the requirement for transmission and high voltage distribution assets

Decentralization of control, DNOs operate local control areas and provide system services with the help of DG

Individual microgrids grids are able to operate autonomously in the case of loss of supply from the higher voltage networks (islanding)

Source: Compiled from Strbac (2006) and Strbac et al. (2007).

EBL

478

Competitive Electricity Markets

•

“Accommodation: DG is accommodated into the current market with the right price signals. Centralised control of the networks remains in place. • Decentralisation: The share of DG increases. Virtual utilities optimise the services of decentralised providers through the use of common communication systems. Monitoring and control by local utilities is still required. • Dispersal: DG takes over the electricity market. Micro-grids and power parks effectively meet their own supply with limited recourse to grid-based electricity. Distribution operates more like a coordinating agent between separate systems rather than controller of the system.” While the third stage seems to be far away, the second stage is already reality in places such as Denmark. Denmark, with its 50% share of DG, is a pioneering country when it comes to implementing and testing new network concepts. The Danish TSO energinet.dk, a state-owned and fully unbundled company, is currently implementing the cell concept, i.e., shifting more responsibility for network control to the DNOs to balance generation and demand locally as far as possible (Lund et al., 2006). 13.2.4. DG and electricity network regulation Chapter 13 focuses on the integration of DG into electricity networks. In the “standard model” of electricity liberalization, it is now widely agreed that in generation and supply, competitive markets can best achieve efﬁcient outcomes. Electricity networks, on the other hand, are unbundled from the competitive business and network tariffs are regulated. The rationale for regulating networks, while opening other parts of the system for competition, is that electricity networks, as opposed to many telecommunication networks, are still natural monopolies. The main objective of network regulation has been to promote short-term efﬁciency. This is achieved through

EBL

•

increasing the efﬁciency of network operation and investment (productive efﬁciency); • ensuring efﬁcient charges for network users, i.e., avoiding monopoly rents (allocative efﬁciency); and • ensuring non-discriminatory charges for all network users in order to promote competition in generation and supply (level playing ﬁeld). In the past, the main focus of network regulation in most liberalized markets has been on short-term efﬁciency, i.e., the reduction of network costs and tariffs. Longer-term issues and network development have only recently come to the fore (e.g., Hirschhausen et al., 2004), partly as a result of a number of electricity blackouts due to network outages. As Burns and Riechmann (2004, p. 211) have pointed out, “while historically there have been concerns about over-investment, there is now a growing unease about under-investment.” This has led to a theoretical and practical debate as to how different regulatory approaches affect investment behavior and how regulation needs to be designed to stimulate efﬁcient investment (Brunekreeft and McDaniel, 2005; for the discussion on generation adequacy see Moran, this volume). Table 13.2 gives an overview of the interaction between different stages of network development and various modes of network regulation. While stage 1 can be associated with the monopolistic era, most systems are currently in stage 2, “Sweating the assets,” where efﬁciency improvements are the main focus. Some systems are moving to stage 3, “Renewing the system,” i.e., securing investment is gaining importance. DG integration

Table 13.2. The interaction between network development and regulation. Building the system

Sweating the assets

Renewing the system

Transforming the system

Main objective

Building the electricity system, security of supply

Cost-reduction (OPEX), short-term efﬁciency

Supply security, efﬁcient investment within the existing system (like-with-like replacement)

Dominant regulatory approach

Rate-of-return regulation

Incentive regulation

Incentive regulation, complemented by targeted incentives, separate treatment of investment?

Main shortcomings

Lack of efﬁciency, not much innovation, overinvestment

Coordination of network and generation

Coordination through vertically integrated monopolies

Static efﬁciency improvements at the expense of investment and innovation Coordination necessary between existing network and new DG plants, price signals

Focus on investment adequacy may hide opportunities for system change Coordination necessary between network and generation development, price signals plus planning

Supporting certain system transformations, e.g., toward sustainability, higher shares of RES/DG Investment need as a window-of-opportunity for structural change Incentive regulation, complemented by targeted incentives and more long-term coordination mechanisms Danger of transformation in “wrong” direction, not keeping options open

Source: Authors.

EBL

Non-market-based mechanisms become more important, e.g., shared vision of the future energy system to coordinate system transformation

Competitive Electricity Markets

480

may be one factor taking us to stage 4, “Transforming the system.” In both stages 3 and 4, network regulation is concerned with replacing existing assets, yet while stage 3 is mainly concerned with like-with-like replacement, stage 4 is about transforming the existing network structure to implement new concepts as outlined above. The following two sections of this chapter analyze the relationship between network regulation and DNOs on the one hand and DNOs and plant operators on the other hand and how DG can be taken into account on both levels. For a more detailed discussion of network regulation to promote network innovation and transformation, that becomes necessary when DG reaches a certain level (see Section 13.2.3), see Bauknecht et al., 2007; Holt, 2005. 13.3. Network Regulation and Problems of DG Integration 13.3.1. DNO incentives under different regulatory regimes This section provides an analysis of the problems for DG that can result from the regulatory framework that is in one way or another applied in most liberalized markets. Lack of incentives for the DNO or even negative incentives resulting from traditional regulation can constitute a major barrier for DG (Connor and Mitchell, 2002). Importantly, any regulatory framework inevitably implies incentives for or against DG. This section analyzes the mechanisms through which DNOs are affected by DG and how this translates into economic costs or beneﬁts for the DNO under different regulatory regimes. DG can have the following effects on the network and the network operator:

EBL

•

DG can cause additional costs, both operational and capital expenditure (Opex and Capex); • DG can entail network beneﬁts; • DG can reduce the volume sold over the network; and • DG can replace or defer network investments. The incentives of the DNO depend on the regulatory framework, speciﬁcally treatment of revenue or proﬁt regulation and unbundling rules. The precise direction of these incentives depends strongly on whether DG causes costs or beneﬁts for the DNO as depicted in Table 13.3. It gives an overview as to how the effects of DG affect the DNO under different regulatory regimes, taking into account two main elements of “network governance”: the degree of unbundling and the mechanism to determine the level of network tariffs. 13.3.1.1. Additional network costs While additional costs may not hurt the DNO under cost-plus approach, assuming that they are accepted by the regulator, the cost-cutting environment that is being established in most liberalized markets through some form of price-based regulation can be detrimental to DG. The basic formulae of regulation through revenue and price caps are

Rt Rt−1 · 1 RPI − X revenue cap and n i=1

pit · Qit−1 Rt−1 · 1 RPI − X tariff basket price cap

Table 13.3. The effect of DG on the DNO under different regulatory regimes. Mechanism to determine the level of network tariffs

Unbundling

Rate-of-return regulation

Price cap

If DG causes additional costs (CAPEX/OPEX)

Maximizing rate-base, “Gold-plating” (Averch-Johnson effect) Incentive for DG, especially if it leads to network expansion

Unbundling may impede system optimization, i.e., can lead to additional costs and make it more difﬁcult to capture network beneﬁts of DG (Brunekreeft and Ehlers, 2006)

If DG causes network beneﬁts

No incentives to reduce costs through DG

If the load on network is reduced through DG (Ackermann, 2004, pp. 186–210) If DG replaces network investments

Lower volume can be offset with higher tariffs to meet proﬁt target

Reduced costs lead to higher proﬁts during the regulatory period incentive against DG Benchmarking gives additional cost-cutting incentives (Mitchell and Connor, 2002): If only OPEX is benchmarked, incentive to shift OPEX to CAPEX (e.g., expand network instead of intelligent network operation) DNO can beneﬁt from reduced costs through DG Long-term beneﬁts may entail short-term costsIncentive regulation can give an incentive to reduce short-term costs at the expense of long-term beneﬁts Lower volume leads to DNO may be able to lower revenue incentive offset volume reduction against DG (RAP, 2000) through a price increase

Reducing or deferring costs can be attractive, but DNO may not like to shift costs from CAPEX to OPEX

If DNO is not allowed to be involved in DG generation, incentive against network replacement. This needs to be compared with the beneﬁts for DNOs due to reduced costs

Source: Authors.

Against DNO’s incentive to maximize rate-base

Revenue cap

EBL

Without unbundling, incentive against volume reduction exacerbated Third-party DG may compete against an integrated ﬁrm’s own generation capacity and thus the integrated DNO may have a disincentive to connect DG

482

Competitive Electricity Markets

There are many variations, but these two will do perfectly for the purpose of this chapter. In the formulae, Rt is the revenue in period t, pi and Qi are the price and quantity of product i (I = 1,…,n). RPI is the retail price index as a measure of general inﬂation, and X is a proxy for the expected productivity increase and determines the tariff or revenue reductions sets by the regulator. Under this type of regulation, the incentives for the DNO are to cut costs. The X-factor is determined by the regulator ex ante for the entire regulatory period. The idea of the regulation is that proﬁts that are decoupled from costs and efﬁciency gains over and above those already anticipated by the regulator and internalized through the X-factor can be retained by the company during a ﬁxed regulatory period. This holds for any level of the X-factor as long as it is not dependent on own past performance. As a result, the DNO under “standard” incentive regulation will want to avoid additional costs caused by DG. There are examples that regulators are unwilling to accept additional costs caused by DG, which obviously sets an incentive against DG. It is sometimes argued that price- and revenue-cap regulations have different effects with regard to DG and “price cap regulation generally discourages distributed resources. Revenue cap regulation does not” (Moskovitz, 2000, p. 3). While this argument can be valid with regard to the volume reduction caused by DG (see Table 13.3), the argument that a revenue cap (as opposed to a price cap) allows the network company to increase network tariffs if DG causes additional costs (Ackermann, 2004, p. 285) does not hold. As revenues are capped, increased costs cannot be offset through higher revenues, at least during a regulatory period. Consequently, both price- and revenue-cap regulations tend to provide an incentive against DG. When analyzing the costs of DG and the way they affect the DNO, it is important to differentiate between different cost categories, namely operational costs (Opex) and capital costs (Capex). Depending on the regulatory treatment of these costs, the DNO may beneﬁt from shifting costs between them, which is likely to inﬂuence its attitude and strategy toward DG. Separate treatment of Opex and Capex tends to focus on incentivizing operating cost efﬁciency, with capital cost allowances established through more traditional utility planning and cost-of-service regulatory accounting methods including the speciﬁcation of a rate base (Joskow, 2006, p. 21). The separate treatment of Opex and Capex is likely to not only promote network investment but also distort overall cost optimization. If the network is upgraded to accommodate additional DG, Capex will increase. Other approaches to network management (e.g., active management, see Section 13.2.3) can reduce the need for network upgrades and hence Capex, but can entail additional Opex, especially if the lower network capacity needs to be managed actively. If Opex is benchmarked and Capex is not, the DNO is likely to be in favor of the more capital-intensive option, even though the overall costs can be higher. Price-based regulation with a focus on Opex reduction can thus have a similar effect on the DNOs choice as rate-of-return regulation with its incentive against capital-saving investments. The separate treatment of Capex and Opex in the United Kingdom is a case in point.4

EBL

4

“Operational expenditure is benchmarked against other DNOs, with the intention of bringing all DNOs nearer to the frontier of the DNO which, taking account of land area and population, has lowest operational expenditure. This, all things being equal, means that the DNO will try and reduce operational expenditure and thus it militates against their undertaking new tasks with no obvious regulated asset base beneﬁts” (Mitchell and Connor, 2002, p. 12.).

Distributed Generation and the Regulation of Electricity Networks

483

13.3.1.2. Network beneﬁts As compared to a rate-of-return regime, incentive regulation in principle increases the incentive for DNOs to utilize DG to reduce network costs. In practice, however, it may be difﬁcult for the DNO to reap the beneﬁts, especially if DG plants do not operate in line with network requirements. The beneﬁts of DG depend on the status of the system where they are to be installed and are very site speciﬁc (see discussion of network costs and beneﬁts above). As long as electricity networks are in good shape and have a high reliability, grid-related beneﬁts of DG tend to be relatively low. While there are likely to be beneﬁts in the long term, in the short term the DNOs tends to incur additional costs (see Section 13.3.1.1). For example, instead of improving system reliability, in the short term, DG may represent an additional challenge to system reliability and therefore require additional measures to compensate this. For the beneﬁts to be reaped, it will be important that DG will not only be connected but also be integrated into network operation (see Section 13.3.2.2). While costs will be incurred simply by connecting DG, beneﬁts will in many cases only become visible if the system adapts. Additional costs may thus occur upfront but cost savings only emerge much later and are uncertain (e.g., because this also depends on overall system development). This is particularly problematic if incentive regulation gives an incentive to reduce short-term costs at the expense of long-term beneﬁts. 13.3.1.3. Load reduction through DG A major concern for network operators with regard to DG is that it reduces the electricity volume and/or the peak load on their network, with negative effects on their revenues. As there are hardly any variable costs in the network, proﬁts would be hurt accordingly. Exactly how the deployment of DG reduces the utilization of the network depends on the speciﬁc case. For an analysis of generic cases, see Ackermann (2004, pp. 186–210). If, especially, DG is installed on the customer side of the meter and provides mainly on-site power production, DNO sales are adversely affected (RAP, 2000). Developments that enable the deployment of micro-generation by small consumers or microgrids serving a speciﬁc area, and where the meter is between the utility network and the microgrid, could exacerbate this problem for the DNO. The further deployment of microgrids depends on both the technical advances and the legal framework in place. In Germany, for example, there have been extensive discussions as to whether local networks have a right to get connected to the medium- rather than the low-voltage network. This would increase their attractiveness and reduce the revenue of the DNO. As long as the network operator is not fully unbundled from the supply business, the integrated company loses not only from reduced network utilization but also from reduced electricity sales of its supply business. While revenue- or price-cap approaches have the same effect with regard to additional costs of DG (see Section 13.3.1.1), the reduced utilization of the network can have different repercussions under the two regimes. Under a price cap, DNOs will try to maximize sales, while under revenue cap it will at least try to meet the volume forecast. If DG reduces volume traded on the network, the following happens: While under a price cap, the network operator is not allowed to offset a volume reduction below the forecast due to DG with higher tariffs, this would in principle be possible under a revenue cap. This gives the company an additional degree of freedom besides cost reduction, although the company may prefer to meet the forecast rather than to increase tariffs.

EBL

484

Competitive Electricity Markets

While it is difﬁcult for the unbundled DNO to inﬂuence electricity demand, DNOs can inﬂuence the additional connection of DG and will be unwilling to support DG if they jeopardize their revenue. 13.3.1.4. DG replacing network investments It is often claimed (e.g., the already mentioned studies by Little, 1999; Iannucci et al., 2003; Gumerman et al., 2003) that one of the main advantages of DG is that it can replace or defer network investment that becomes necessary either because old network equipment needs to be replaced or because demand increases in a network area and can no longer be covered by the existing capacity. Instead of building additional network capacity, a DG plant could provide the generation capacity locally. As opposed to the volume reduction that has just been discussed, this effect of DG can result from all DG plants, not just plants on the customer side of the meter. At the same time, it does not happen automatically as generation from DG goes online, but only if these plants meet certain requirements (see Section 13.3.2). Although this is clearly a potential network-related beneﬁt of DG, the DNO may be opposed to it. The main reason for this is that replacing the network with DG also means replacing the DNO’s business with the generator’s business and introduces a form of competition to the network. For the DNO, this will be particularly problematic if (a) the DNO is unbundled from other functions, can therefore not invest in DG itself and has to leave this up to third-party generators (Brunekreeft and Ehlers, 2006); and (b) if DG replaces distribution network assets rather than transmission system assets operated by a separate transmission company.

EBL

While under rate-of-return regulation the DNO would not want to replace his/her own asset base with other companies’ assets, under incentive regulation combined with benchmarking a DNO can stand to beneﬁt from deferring network investments, especially if DG plants are located in high-cost areas of the network. As increasing DG to replace network assets would reduce CAPEX but increase OPEX costs, the incentives of the DNO also depend on whether these cost categories are treated in a way that allows for overall optimization. What is important to note is that while DG may replace network capacity and thus reduce the DNO’s business, it can also entail new tasks for the DNO. An increasing share of DG requires them to be integrated into network operation, which presupposes that the local distributors are included in the control and regulating work and will be responsible for adjusting the local production to demand (see above: cell concept in Denmark). Moving from passive to active networks thus can also mean a shift from “passive” to “active” DNO with new business opportunities, provided the regulatory framework allows for such a development.

13.3.2. Coordination problems between network operators and DG Electricity supply requires a close coordination of different stages in the value chain. As electricity cannot be stored, generation must exactly match electricity demand in real time, which requires appropriate operational mechanisms. There also needs to be enough capacity available to meet peak demand, which requires timely and sufﬁcient investment. Furthermore, a close coordination between network and generation is required, in terms

Distributed Generation and the Regulation of Electricity Networks

485

of both investment and operation, in order to provide a reliable and efﬁcient electricity supply. Liberalization has opened the market for new entrants and is based on unbundling the competitive and non-competitive functions of the electricity market, namely generation, network, and supply. Therefore, liberalization leads to a new set of externalities and coordination problems, which require new institutional approaches. As a result, alternative modes of coordination between different actors and unbundled functions operated by separate companies are required to make a more decentralized system work and optimize the overall beneﬁts of DG, i.e., both beneﬁts to the plant operator and the DNO. Coordination difﬁculties are not typical for just DG and distribution networks, but they tend to be more severe in this context because of the following reasons: • •

•

• • •

Distribution systems are generally not equipped to deal with generation (see Section 13.2.2). In predominantly centralized systems, DNOs do not have much experience with generation connected to their network and operate under planning and operational routines that do not normally take into account DG. As opposed to the high-voltage network where the number of plants and actors is relatively small, in the case of distribution networks, the number of actors involved is likely to be signiﬁcantly higher and plant projects are both small scale and more numerous, per deﬁnition (Brunekreeft and Ehlers, 2006). DG plants often operate within special regimes to support certain generation technologies. DG plants often generate intermittent power or provide both heat and electricity which tends to make coordination more difﬁcult. As a rule, new generation capacity causes costs somewhere in the network to facilitate new power ﬂows; in other words, new generation capacity can have “deep” network impact. The costs could be attributed to the new generation assets for big projects, but deep connection charging turns out to be a substantial barrier to small DG. The practical approach to small DG thus is shallow connection charging, leaving the network effects to the DNO, creating external effects.

EBL

The following will brieﬂy describe coordination problems between distribution networks and distributed generation in terms of network planning, DG connection, and DG operation. Additionally, a coordination problem can arise from the need to develop innovations, including DG plants that are better prepared to provide network services, interfaces between plants and networks, and new network concepts. Innovations carried out in different areas by different actors need to be compatible with each other, and there is a need for standardization between DG plants, networks, and markets. Plant technology providers, for example, need to know what DNOs require from DG plants to improve system integration. 13.3.2.1. DG connection and network planning The ﬁrst coordination problem arises in the investment phase, in terms of both siting of individuals DG plants and overall network development. Siting can heavily inﬂuence the grid-related costs or beneﬁts. For a DNO, a certain plant location may reduce grid congestions, whereas another location may pose grid capacity problems. For the plant operator, siting may also be crucial because of local heat demand to be covered by CHP plants or availability of energy resources such as wind. A DG plant

486

Competitive Electricity Markets

should be connected to the grid at a point where the costs for upgrading the grid are minimized (subject to the site-speciﬁc costs incurred by the DG operator) or the beneﬁts are maximized, e.g., if a plant can replace new cables. The siting problem has a horizontal and a vertical dimension, i.e., concerns both the location in a given network level and the voltage level. Concerning the latter, there are conﬂicting objectives if the DNO wants to connect plants to a voltage as high as possible in order to minimize the network impact of DG, whereas the DG operator favors a lower voltage level to reduce connections costs. As for DG replacing network capacity, there are a number of issues that need to be resolved, including the investment and operation mode of the plants (for the operation mode, see below). For DG plants to be able to replace network investment, they must obviously be located in the network area where demand exceeds network capacity. However, as soon as the DG capacity installed exceeds the local demand, the distribution may become an exporter of power and network capacity may start to be driven by DG rather than by demand customers. If network problems are resolved with non-network assets (DG plants) in an unbundled world, the question arises who is in charge of system security and reliability. This has a short-term and a long-term dimension. In the long term, coordination problems can result from the fact that the two actors tend to have different investment horizons. While the DNO may plan to upgrade the grid to provide security of supply for up to 40 years, a DG plant that can replace such an investment by providing power locally, may go off-line earlier. In any case, because the DNO is in charge of grid security, it may not like the idea of relying on a DG operator to provide that security. Beyond the connection of individual plants, network planning expansion must be in line with the development of DG. This goes beyond the optimal connection of individual plants to the existing grid but refers to the overall development of DG capacity and the corresponding planning of the grid. The expansion of the grid depends not just on the location of individual plants but, for example, also on density of plants, with a higher density generally leading to higher costs.5 The resulting coordination problem is well illustrated by the following quote from New Zealand (Bertram, 2006, p. 230): “Since 2002 a rush by large incumbent generators to build wind farms is raising a raft of difﬁcult coordination problems, since the location of favourable sites for wind farms, and of the hydro generation assets that will be used to back-up wind generators, does not always coincide with the existing grid infrastructure, presenting the grid’s operator and the new regulator with investment and coordination requirements not foreseen even a few years ago.” Presently, in most countries, DG is not incorporated into grid planning methodologies (de Jong, 2006), and there is no joint planning of DG and grid development. The DNO can pursue a network development strategy that is robust, ﬂexible, or focused (de Vries, 2004, pp. 216–7). Robust strategies are expensive as they are designed to accommodate various generation scenarios. A ﬂexible strategy is able to quickly adapt to new market developments. However, it is often not available for networks with huge sunk costs and long asset life times. A focused strategy develops the network in a way that it is able to deal with a rather narrow band of generation development. This is cheaper, yet bears

EBL

5 For example, Cao et al. (2006): “Greater installed capacity of DG and denser DG penetration will result in more feeders with voltage rise problems in rural networks and will require that more protection assets are updated in urban networks. Hence the reinforcement cost will increase as the DG penetration level increases.”

Distributed Generation and the Regulation of Electricity Networks

487

the risk that if the development of generation turns out to be different from what was expected, the network may not able to deal with it. Liberalization and unbundling have made a focused strategy more difﬁcult. DG exacerbates this further, as decisions to build DG plants can be made by a large number of actors, which are difﬁcult to coordinate. The exacerbated coordination and externality problem enhances the role of efﬁcient network pricing to signal investment needs. Locational marginal pricing (nodal spot pricing) would, although imperfectly, go a long way in doing just this. 13.3.2.2. DG and network operation In operational terms, the requirements of the network on the one hand and the plant operator on the other hand may also differ. What is best from a network point of view may be suboptimal for the plant operator and vice versa. DG plants are generally operated according to price signals from the electricity market or a special regime, e.g., a feed-in system as in Germany that makes them maximize the total annual kWh output. In both cases, DG operators do not take into account how their generation corresponds to the requirements of the distribution network. The most important point is the contribution of a DG plant to peak load on the network. DG plants may either reduce the load needed from the transmission system or increase exports to the transmission system, depending on its generation proﬁle. The effect of DG plants on the voltage level on the network is also relevant (Jenkins et al., 2000). While under current operational regimes plants go off-line if there is a voltage problem the network cannot deal with, network operators cannot inﬂuence the generation proﬁle more gradually, i.e., only partly reduce generation or ask the generator to operate during peak hours. The coordination requirement increases signiﬁcantly if lines are to be replaced by DG plants. In this case, the DNO, who is still in charge of network security, needs to rely on the contribution of the DG plant, especially during peak load.

EBL

13.4. Regulatory Approaches to DG Having analyzed the main problems with regard to DG, this section presents possible solutions. Again, this will be done in two steps. It ﬁrst looks at how DNOs can be given incentives to connect DG. Second, it discusses approaches to coordinate networks and DG in an unbundled system.

13.4.1. Designing DG incentives for DNOs Section 13.3.1 discussed four different effects of DG. This section will focus on how additional costs of DG that are likely to prevail in the short term can be dealt with in the regulatory framework.6 Assuming that a network operator with a higher share of DG 6 As for the other aspects, we consider the volume reduction less of a problem, at least as long most DG plants are connected directly to the DNO’s network. We have already discussed that a revenue cap is, in principle, superior to a price cap to deal with this issue. Other options discussed are compensation for volume reductions, e.g., through regulatory accounts or multiple-driver cap schemes (Bundesnetzagentur, 2006; Takahashi et al., 2005; Thomas et al., 2001).

488

Competitive Electricity Markets

connected to its system has to bear higher short- or mid-term costs, the question arises how these additional costs can be taken into account. Three principles should apply: •

The DNO should not be penalized if there is a lot of DG connected to the network, leading to higher network costs, for example, because there is a lot of primary energy potential for renewables in the system. Otherwise, DNOs will be opposed to DG. DG-related costs should therefore explicitly be recognized in the regulation of allowed network tariffs/revenues. • The DNO should connect a given amount of DG as efﬁciently as possible. While the number and type of DG connections in a network area are exogenous to the DNO,7 the DNO may well have an inﬂuence on the costs of individual connections and reinforcements of the network that become necessary due to DG. • Finally, DNOs should have an opportunity to earn an extra proﬁt from DG connections. As Edison Electric Institute and NERA (2006) have put it, utility involvement in distributed resources cannot be optimized or maximized merely by the removal of disincentives or barriers. This removal is a necessary step, but it is not sufﬁcient. To elicit the commitment of time, energy, and creativity needed to fully exploit the potential for efﬁcient distributed resource development, utilities must be allowed to make a business out of DR. Positive incentives seem necessary to overcome path dependencies of DNOs and help them evolve from organizations whose planning routines and know-how are geared toward running the passive end of central systems to “active DNOs” that facilitate optimal network development, taking into account distributed resources. Giving companies an opportunity to earn extra proﬁts if they achieve certain targets deﬁned by the regulator (and policy) is in line with the general mechanism of incentive regulation, where companies are normally incentivized to increase productive efﬁciency at the expense of allocative efﬁciency. Given these objectives, it is clear that DG-related network costs should be included in the calculation of allowed revenues or tariffs, but a DNO should not simply be reimbursed for all the costs it declares. A balance must be struck between the connection incentive and the efﬁciency incentive. At the same time, positive incentives should be linked to the efﬁciency of DG connections. The following paragraphs will discuss various approaches to take into account the costs of DG in the network regulation depicted in Fig. 13.4:

EBL

• • •

full cost pass-through, standardized revenue driver (e.g., per kW of DG connected), a combination of a (partial) pass-through and a (supplementary) per kW revenue driver, • ﬁnally, a combination of these mechanisms with a menu of sliding scales will be proposed. Such a self-selection incentive scheme can work quite well if the regulator does not know the costs of network investment and does not want to frustrate new investment. 7

This may be changed through joint planning between DNOs and plant operators. Also, DNOs are in a strong position to prevent DG connections. But there are numerous factors inﬂuencing DG development that cannot be controlled by the DNO.

Distributed Generation and the Regulation of Electricity Networks

489

Full pass-through

DNO revenue per kW

Percentage passthrough + supplementary revenue driver per kW revenue driver

Percentage cost pass-trough Supplementary per kW revenue driver

DNO costs per kW Average costs of DG connection Fig. 13.4. The effect of pass-throughs and per kW revenue drivers on revenues. Source: Authors.

EBL 13.4.1.1. Full cost pass-through A well-known mechanism within incentive regulation is the so-called z-factor (for a brief overview, see RAP, 2000). This is an adjustment factor outside the incentive regulation mechanism used to deal with extraordinary costs beyond the DNO’s control, such as costs resulting from changes in laws, taxes, or extreme weather conditions. While under traditional cost-plus regulation, such costs would simply be taken into account during or possibly trigger the next rate case, under incentive regulation, regulatory periods are ﬁxed and costs that cannot be controlled by the DNO should not be subject to the RPI-X incentive mechanism. The z-factor is a mechanism to allocate risk between DNOs and customers. If a 100% z-factor is applied, risk is fully passed to customers. Under a full cost pass-through mechanism, the revenue of DNOs would be directly linked to the costs incurred. There are different ways to implement the z-factor. Costs could be recovered through a z-factor based on a forecast made at the beginning of each regulatory period and corrected at the end, ex post on an annual basis, at the moment the cost is incurred or at the following regulatory review. The z-factor is added to the revenue cap (or price cap) as an additional element which is not subject to the RPI-X mechanism: Rt ≤ Rt−1 · 1 + RPI − X + z where z represents (100% of the costs of) the DG connections.

490

Competitive Electricity Markets

In this case, the assumption is that DNOs cannot inﬂuence DG connection costs. This is only partly true8 and a full cost pass-through does not incentivize the DNO to look for more efﬁcient connection options. If DNOs can simply pass through any DG-related costs they incur, they do not carry any risk. They, therefore, would lack the incentive for efﬁcient DG connections and will be tempted to game. They will try and label costs that are not to do with DG as DG-related costs and shift costs from the general RPI-X incentive mechanism into the special DG regime. A full-cost pass-through, therefore, does not seem to be appropriate. 13.4.1.2. Volume-related revenue driver Instead of including the costs actually declared by the DNO into the calculation of allowed revenues, the regulator could set a ﬁxed volume-related revenue driver, i.e., the allowed revenue would increase depending on the amount of kW connected. In principle, this is a price- or a revenue-cap mechanism. The revenue of the DNO would thus depend not on the costs but on the DG capacity connected, thereby incentivizing the DNO both to connect DG and to reduce costs. The revenue driver would be included in the revenue formula just like the cost pass-through element: Rt ≤ Rt−1 · 1 + RPI − X + y · kWDG where y is an average compensation for each kW connected DG set by the regulator. This revenue driver could, for example, be calibrated to reﬂect the average incremental costs of DG connections. The more the costs of a DG connection remains below the revenue driver level, the higher will be the additional proﬁt of the DNO. Consequently, there is a positive incentive linked to the costs of the DG connection. Above the revenue driver level, the DNO would lose money from connecting DG. DNOs could be given an additional positive incentive through increasing the revenue driver, yet this would obviously happen at the expense of (allocative) efﬁciency. Under such a scheme, the DNO will try to reduce connection costs below the revenue driver value to earn an extra proﬁt. This can be done by the following:

EBL

•

Improving the efﬁciency of connecting a given DG project to achieve below-average costs. This would be a positive efﬁciency improvement. • “Cherry-picking,” i.e., preventing DG projects with above-average connection costs where possible. This would be negative, especially if the overall beneﬁts of such a project exceed its costs despite above-average connection costs. • Improving coordination with DG operators, i.e., through labeling and making the network costs at various connection points more transparent, in other words, market the network and improve coordination with DG operators. This refers to Section 13.4.2. While under the full pass-through mechanism it may be difﬁcult for the regulator to determine the costs that have been caused by DG and can be passed through, it will be just as difﬁcult to establish an appropriate level for the volume-related revenue driver. If it is too low, the regulation will frustrate network investment to facilitate potentially 8 For some DNOs, it will be easier to connect DG plants than for others, e.g., because the DG potential in a network area is more in line with the existing network structure. Irrespective of that, the DNO can inﬂuence how efﬁciently a given DG potential is connected to the network and whether intelligent solutions like active management are chosen where this is more cost-effective.

Distributed Generation and the Regulation of Electricity Networks

491

beneﬁcial new DG. It is even more challenging to design a more sophisticated revenue driver that does not simply use kW as a cost driver but is differentiated by DG type, size, or location, especially as there is not much experience yet with DG-related cost drivers. 13.4.1.3. Partial cost pass-through plus volume-related revenue driver A third option would be to combine a partial cost pass-through with a volume-related revenue driver. This allows the development of an incentive scheme for DNOs which balances different incentives, provided the regulator has a good understanding of the network costs of DG. This is difﬁcult, especially as DG costs are very case sensitive and DG is a relatively new issue. As can be seen from Fig. 13.4, this mixed mechanism can be designed so that the DNOs’ additional revenue allowance exceeds their additional costs as long as the cost of a DG connection remains below the average (or some other value). If a DG connection costs more than the average, the DNO would lose money. At the same time, the additional revenue DNOs can achieve is more ﬁne-tuned compared to the one-size ﬁts all stand-alone revenue driver described above.9 An example for the explicit treatment of DG-related costs in DNO regulation is the approach employed in the United Kingdom under the so-called hybrid incentive for DG-related distribution costs covering both investment and operation and maintenance. Hybrid refers to a mixture of a partial cost pass-through and a volume-related per kW revenue driver. According to Ofgem (2004), the hybrid mechanism is to “combine incentives for efﬁciency with protection against cost uncertainty.” The UK’s hybrid incentive comprises the following elements: • • • •

EBL

a pass-through rate of 80%; a supplementary incentive of £ 1.5 per year for each kW of DG capacity connected; an additional £ 1 per year for operation and maintenance; a cap (two times the cost of capital) and a ﬂoor (cost of debt) for overall returns; and • incremental unit costs above £ 200/kW are paid by the plant operator through connection charges. This enables DG projects to be realized, which have high networkrelated costs, but can be justiﬁed by the economics of the plant itself. The UK’s hybrid incentive may be characterized as a ﬁrst attempt to overcome the problems of simple cost pass-throughs, while balancing various objectives and incentives. Although in theory DNOs can earn additional revenues through this mechanism, in practice, the mechanism does not seem very effective yet. The hybrid incentive will be reviewed as part of the 2010 distribution price control review. 13.4.1.4. DG in the RPI-X mechanism and benchmarking If the costs of DG are not dealt with outside the RPI-X incentive mechanism, for example, through one of the mechanisms described above, the RPI-X mechanism itself needs to 9 Such a mechanism is not an additional incentive for the connection of DG but a mechanism to recover the costs of DG, which may be designed to give additional incentives. It should therefore not be confused with a targeted incentive, whereby a company gets rewarded for achieving certain predeﬁned performance target such as increasing the connection of DG. As Joskow and Schmalensee (1986) have pointed out, a targeted incentive scheme can lead to distortions, if companies are rewarded for achieving performance targets, the costs of which are recovered through a separate mechanism.

492

Competitive Electricity Markets

take into account the costs of DG. As it is the main objective of incentive regulation to increase the efﬁciency of networks, this objective needs to be reconciled with the additional costs DG plants can entail. In other words, the analysis of a DNO’s efﬁciency and its potential to increase its efﬁciency must incorporate the costs of DG. As most regulators are implementing some form of comparative efﬁciency analysis (benchmarking), DG would need to be included in the benchmarking mechanism as a cost driver for the distribution network. For this to work, it needs to be understood what drives the network-related costs of DG, e.g., size, type. There are hardly any practical experiences yet with including DG into benchmarking exercises. The new German regulator Bundesnetzagentur (Federal Network Agency), in its report from June 2006 which set out the new incentive regulation mechanism to start in 2009 (Bundesnetzagentur, 2006, p. 238), has listed DG as a potential external cost driver to be considered in the efﬁciency analysis. Perhaps surprisingly, ﬁrst estimates of efﬁciency effects of DG suggested no statistical signiﬁcance. 13.4.1.5. Menu of sliding scales Combining the elements of the schemes described above leads to a self-selection scheme, i.e., a menu of sliding scales. Key to self-selection incentive mechanisms is an information asymmetry between principal and agent. If the principal does not know the costs of the agent, the principal can design an incentive compatible mechanism which triggers the agent to reveal the costs truthfully. The phenomenon is quite well known from the formal literature but has recently found an application in the United Kingdom with Ofgem’s 2005 distribution price control. Both have been described by Joskow (2006). The same ideas can be applied to DG regulation. We have argued above that the information asymmetry of the cost effect of DG on the DNO is very substantial as the cost effect is strongly case sensitive. The following analysis limits the attention to DG only, assuming all other costs have been taken care of. Let z be the network cost of new DG (with beneﬁts, z would be negative). Denote y as a price cap, which is the average revenue for a kW-connected DG. The price cap from this is y·Q, where Q is the quantity of connected DG. Let b be a parameter which determines the sliding scale, being a combination of cost pass-through and a price cap. The allowed revenue R is

EBL

R = b · y · Q + 1 − b · z

(1)

As can be seen, this combines the two elements above. The sharing parameter b is a function of the inherent network costs of connecting DG, denoted by : b(). The regulator does not know , which, if at all, is only known by the DNO. If b = 0, the regulation is full cost pass-through. With b = 1, regulation is fully price-cap. A sliding scale is a sharing parameter between the extremes: 0 < b < 1. With for instance b = 0.5, 50% of the reported costs of 1 kW-connected DG can be passed through, and the DNO receives 50% of y for this 1 kW-connected DG. The sliding scale mechanism turns into a self-selection mechanism if the ﬁrms can choose b themselves, given the values for y set by the regulator. In fact, to make this incentive mechanism work, y should be an increasing function of b. This means that if the ﬁrm chooses a high b (low cost pass-through), the price cap will be relatively high and reverse. Appropriate choice of y(b) by the regulator provides that the incentive mechanism is incentive compatible, and thus the choice of b by the ﬁrms will be optimal. Note that the choice of y(b) should fulﬁl the incentive compatibility constraint and does not necessarily

Distributed Generation and the Regulation of Electricity Networks

493

bear a relation with costs. Firms will select themselves and thereby reveal themselves as high-cost ( is high) or low-cost ﬁrms ( is low). This means that a DNO, which knows that DG connected to its network causes relatively high network costs (high ), will select a low b, and thereby a low cap and high-cost pass-through. A DNO knowing that its DG is low costs or even beneﬁts the network (low ) will select high b and thereby a high cap (y) and low-cost pass-through. In effect, the incentive mechanism selects the best option in each scenario. If in a network area there is potential of low-cost DG or the DNO has potential to connect DG efﬁciently, this DNO will choose a price cap, thereby it self-selects the incentives to be efﬁcient and connect low-cost DG and the mechanism promotes low-cost DG. The DNO will try to avoid high-cost DG, but by the level of the price cap, this is implicitly desired by the regulator. Also, note that in this case DG, which actually beneﬁts the network, is not frustrated. If, on the other hand, in a network area, there is high-cost DG or the DNO knows it will not be able to connect DG at low costs, it will opt for cost pass-through, thereby the DNO self-selects the incentives to connect high-cost DG because it can pass through the costs. Hence, desired DG is not frustrated by a too low cap. At the same time, DG that beneﬁts the network will not be frustrated because a DNO with predominant network beneﬁting DG would not choose the cost pass-through option. It should be noted that the scheme must be designed carefully to be incentive compatible. It can be done though, as suggested by the example in the United Kingdom, with the 2005 distribution price control.

EBL

13.4.2. Improving coordination between DNO and DG The coordination problem between investment in network and generation plant in a liberalized, decentralized world is potentially severe. DG exacerbates the problem because the DNO (or, the invisible hand) has to deal with many small partners instead of a few big ones. Following Brunekreeft and Ehlers (2006), two different types of coordination problems can be distinguished: external effects, which are not fully internalized, on the one hand and lack of information on the other hand. DG has a “deep” network impact. As indicated above, new DG has costs and beneﬁts for the network. These costs and beneﬁts are strongly case sensitive. Moreover, the costs and beneﬁts are mutually interdependent for different DG investments. An assessment of the costs and beneﬁts of new DG requires case-by-case analysis. With the help of engineering models and analytical cost models this can be done. Once the effect of new DG on the network is known, including factors such as changed power ﬂows, higher or reduced losses, required upgrades, or network deferrals, the costs and beneﬁts can, in principle, be passed on to the new connection. Passing on full costs and beneﬁts of new generation investment on the network to the investor is called deep charging. Deep charging internalizes external effects and thereby improves efﬁcient investment signals. However, deep charging is not feasible in practice, and, if anything, there is a trend toward the opposite (shallow charging). Shallow charging means that a new generation plant pays for the connection to the network but is not charged for any costs, nor compensated for any beneﬁts beyond the connection point. Deep charging contains a number of problems: •

Deep charging implies different charges for each new DG as cost impact differs. As these may be competitors, deep charging is likely to be discriminatory; even if charging would be cost-reﬂective, it would distort the level playing ﬁeld between different new DGs.

494

Competitive Electricity Markets

•

Deep charging implies negotiation between DNO and generator, which is particularly problematic for small DGs. Small DGs typically suffer from an informational disadvantage compared to the DNO. The DNO has the expertise and the scale effects to do the calculations and subsequent negotiations, but more importantly, the DNO is the only one with full information of the power ﬂows. • Lastly, engaging in case-by-case negotiations entails signiﬁcant and real costs, which may be prohibitively high for small DGs. On balance, it can be concluded that deep charging may have the potential to reﬂect costs and beneﬁts efﬁciently but that broadly deﬁned “transaction costs” are simply too high for small DG. Deep charging is therefore not an attractive option if DG connection is to be promoted. Shallow charging does not internalize the external costs and beneﬁts and thereby creates an investment coordination problem. Investment of new DG will be affected in at least two ways. There may be too much or too little new DG, which, as explained above, depends on the incentives of the DNO, which in turn depend on the cost effect and the type of network regulation. In any case, negotiations create the potential to exploit the information asymmetry. New DG may be invested at the wrong location because locational costs and beneﬁts are not reﬂected. Nodal spot pricing, which internalizes these locational effects at least partly, can potentially mitigate this problem. One can imagine the following problem. Suppose that it would be more cost-effective to build new generation capacity at point A than to strengthen the transmission line between A and B. Now suppose that siting for big plants is a problem at A. High nodal prices at node A gives the investment signals to DG investors who might not have a siting problem. Hence, scarcity of generation sites would be reﬂected by nodal pricing to the advantage of DG. Apart from the external effects problem, a genuine coordination problem exists. The investment problem is a simultaneous optimization problem combining new generation capacity and network expansion and upgrade. Typically, piecemeal network expansion leads to inefﬁciencies. For example, Baldick and Kahn (1993) provide illustrative analysis. The investment problem itself is well known; the point to be stressed here is that in a centralized world, the coordination problem is internal to a ﬁrm and in a decentralized world coordination should be solved by market mechanism or new institutions. If the problem is “merely” a lack of information, then presumably it can be solved satisfactorily by designing some centralized information exchange mechanism. The problem is deeper if there is an incentive problem. Within the ﬁrms, the incentive structures between agents is likely to be more in line than among different ﬁrms. The incentive structure thereby makes information credible. In other words, because the incentive structure in a decentralized world may conﬂict, an information exchange mechanism may fail to work because it is not credible. In more practical terms, potential investors in competitive DG will provide information strategically, making planning for other potential investors and the DNO difﬁcult.

EBL

13.5. Conclusions If DG is to move out of its present niche, it needs to be better integrated into the electricity network. DG can entail both network costs and beneﬁts. The case-speciﬁc impact of new DG lies at the heart of Chapter 2‘s main conclusions. First, the case-speciﬁc costs of DG are difﬁcult to understand for the regulator, which makes it difﬁcult to choose between cost pass-through and price capping. The contribution

Distributed Generation and the Regulation of Electricity Networks

495

therefore proposes to further examine the application of a menu of sliding scales to the regulation of DG costs. Secondly, due to the case-speciﬁc network impact of new DG, unbundling of network and generation can lead to coordination problems. First, the network impact of DG is an externality. As deep connection charging is not feasible, it is unlikely to be fully internalized. It is unclear how to address the problem. Second, network planning is a forward-looking process, which requires information about new generation and load. Whereas network operators may have a good notion of the potential for large-scale plants, for DG the picture changes drastically, as DG is decentralized and small scale. While this chapter has focused on incentives for DNOs to connect DG and locational signals for DG operators, more work is required on integrating DG into network operation and incentives for DNOs to implement innovative network concepts. References Ackermann, T. (2004). Distributed resources in a re-regulated market environment. Thesis, Stockholm. Alanne, K. and Saari, A. (2006). Distributed energy generation and sustainable development. Renew. Sust. Energy Rev., 10(6), 539–58. Arthur D. Little Inc. Distributed Generation: Understanding the Economics. Cambridge, MA. Baldick, R. and Kahn, E. (1993). Network costs and the regulation of wholesale competition in electric power. J. Regul. Econ., 5(4), 367–84. Bach, P., John, E.N., Markus, H., Søren, V., and Christian, G. (2003). Active Networks as a tool to integrate large amounts of distributed generation. Paper presented at the conference. Energy Technologies for post Kyoto Targets in the Medium Term, Risø: Roskilde. 19–21 May. Bauknecht, D., Leprich, U., Späth, P., et al. (2007). Regulating innovation and innovating regulation (DG-GRID project report). Bertram, G. (2006). Restructuring the New Zealand electricity sector 1984–2005. In Electricity Market Reform. An International Perspective (Sioshansi, F.P. and Pfaffenberger, W., eds). Oxford: Elsevier, pp. 203–34. Brunekreeft, G. and Ehlers, E. (2006). Ownership unbundling of electricity distribution networks and distributed generation. Compet. Regul. Netw. Ind. 1(1), 63–86. Brunekreeft, G. and McDaniel, T. (2005). Policy uncertainty and supply adequacy in electric power markets. Oxf. Rev. Econ. Pol., 21(1), 111–27. Bundesnetzagentur (ed.) (2006). Bericht der Bundesnetzagentur nach § 112a EnWG zur Einführung der Anreizregulierung nach § 21a EnWG. Bonn. Burges, K. and Twele, J. (2005). Power systems operation with high penetration of renewable energy – the German case. Presented at the Future Power Systems Conference 2005, 17 November 2005, Amsterdam. Burns, P. and Riechmann, C. (2004). Regulatory instruments and investment behaviour. Util. Pol., 12(4), 211–19. Cao, D.M., Pudjianto, D., Grenard, S., et al. (2006). Costs and beneﬁts of DG connections to grid system (DG-GRID project report). Cardell, J. and Tabors, R. (1997). Operation and control in a competitive market: distributed generation in a restructured industry. Energy J. (Distributed Resources Special Issue), 187–210. Coll-Mayor, D., Paget, M., and Lightner, E. (2006). Future intelligent power grids: analysis of the vision in the European Union and the United States. Energy Pol. 35(4), 2453–65. Connor, P. and Mitchell, C. (2002). A review of four European regulatory systems and their impacts on the deployment of distributes generation (Sustelnet project report). de Jong, J. (2006). Review of current utility planning approaches for DG and detailed policy guidelines for network planners to encourage the consideration of DG as an alternative to network infrastructure upgrade (ELEP (European Local Electricity production) project report). de Vries, L.J. (2004). Securing the public interest in electricity generation markets; the myths of the invisible hand and the copper plate. Thesis, Delft.

EBL

496

Competitive Electricity Markets

DoE (U.S. Department of Energy) (2007). The potential beneﬁts of distributed generation and raterelated issues that may impede their expansion. EA Technology Ltd (2001). Likely changes to network design as a result of signiﬁcant embedded generation. Edison Electric Institute, NERA (2006). Distributed Resources: Incentives. European Commission (2003). New ERA for electricity in Europe – distributed generation: key issues, challenges and proposed solutions. Brussels. Gumerman, E.Z., Bharvirkar, R.R., Hamachi LaCommare, K., and Marnay, C. (2003). Evaluation framework and tools for distributed energy resources. Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, CA. http://eetd.lbl.gov/EA/EMP/emp-pubsall.html. Hirschhausen, C., Beckers, T., and Brenck, A. (2004). Infrastructure regulation and investment for the long-term – an introduction. Util. Pol., 12(4), 203–10. Holt, D. (2005). Where has the innovation gone? R&D in UK utility regulation. Oxera, Agenda November 2005. http://www.oxera.com. Iannucci, J.J., Cibulka, L., Eyer, J.M., and Pupp, R.L. (2003). DER beneﬁt analysis studies: ﬁnal report. National Renewable Energy Laboratory, Golden, CO. IEA (International Energy Agency) (2002). Distributed Generation in Liberalized Electricity Markets. Paris. IZT et al. (2004). EurEnDel. Technology and social visions for Europe’s energy future. A Europe-wide Delphi study. Final report. Jenkins, N., Allan, R., Crossley, P., et al. (2000). Embedded Generation. IEE Press. London. Joskow, P.L. (2006). Incentive Regulation in Theory and Practice: Electricity Distribution and Transmission Networks. Cambridge, MA: MIT Press. Joskow, P.L. and Schmalensee, R. (1986). Incentive regulation for electric utilities. Yale J. Regul., 4(1), 1–49. Lund, P., Cherian, S., and Ackermann, T. (2006). A cell controller for autonomous operation of a 60 KV distribution area. Int. J. Dist. Energy Resour., 2(2), 83–100. Mitchell, C. and Connor, P.M. (2002). Review of Current Electricity Policy and Regulation – UK Case Study (Sustelnet projet report). Moskovitz, D. (The Regulatory Assistance Project) (2000). Proﬁts and Progress Through Distributed Resources. Maine. Ofgem (2002). Distributed Generation – “The Way Forward”. Ofgem (2004). Electricity distribution price control review. Appendix – Further details on the incentive schemes for distributed generation, innovation funding and registered power zones. Pepermans, G., Driesen, J., Haeseldonckx, D., et al. (2005). Distributed generation: deﬁnition, beneﬁts and issues. Energy Pol., 33(6), 787–98. RAP (The Regulatory Assistance Project) (2000). Performance-based regulation for distribution utilities. Strbac, G. (2006). Overview of innovative network concepts. Presented at the Conference on Integration of Distributed Generation into Electricity Networks, 8 March 2006, Berlin. Strbac, G., Jenkins, N., Green, T., and Pudjianto, D. (2007). Review of innovative network concepts (DG-GRID project report). Sussex Energy Group, SPRU (Science and Technology Policy Research) (2006). Response to the UK Government’s 2006 Energy Review. Brighton. Swisher, J.N. (2001). Cleaner energy, greener proﬁts: fuel cells as cost-effective distributed energy resources. RMI Sol., 17(3), 10–22. Swisher, J.N. (2002). Small is proﬁtable: The economic beneﬁts of distributed generation. Cogenerat. On-Site Power Prod., 3(4), 17–27. Takahashi, K., Baker, S., and Kurdgelashvili, L. (2005). Policy Options to Support Distributed Resources. Center for Energy and Environmental Policy, University of Delaware, Newark. Thomas, S. et al. (2001). DSM pilot actions, DSM bidding and removal of DSM disincentives from price regulation. A joint project in Italy, Germany and Austria. Part III: price regulation and removal of DSM disincentives in monopoly segments of restructured electricity markets. Final report. van der Vleuten, E. and Raven, R. (2006). Lock-in and change: distributed generation in Denmark in a long-term perspective. Energy Pol., 34(18), 3739–48.

EBL

Distributed Generation and the Regulation of Electricity Networks

497

van Overbeeke, F. and Roberts, V. (2002). Active networks as facilitators for embedded generation. Cogenerat. On-Site Power Prod., 3(2), 37–42. Varming, S., Gaardestrup, C., and Nielsen, J.E. (2002). Review of technical options and constraints for integration of distributed generation in electricity networks (Sustelnet project report). WADE (World Alliance for Decentralized Energy) (2005). World Survey of Decentralized Energy 2005. WADE (World Alliance for Decentralized Energy) (2006). World Survey of Decentralized Energy 2006.

EBL

EBL

This page intentionally left blank

Chapter 14 Global Climate Change and the Electric Power Industry ANDREW FORD School of Earth and Environmental Sciences, Washington State University, Pullman, WA, USA

Summary The warming of the atmosphere and the oceans has been attributed to the release of carbon dioxide (CO2 ) and other greenhouse gasses (GHG) to the atmosphere. CO2 is the principal greenhouse gas, and the generation of electric power accounts for an important share of the CO2 emissions. Scientists and policymakers are calling for major reductions in CO2 emissions and debating a variety of policies to achieve the reductions. The policy options include a combination of targets, regulations, and ﬁnancial incentives to “put a price on carbon.” That price could come from either a carbon tax or a carbon market, but policymakers are inclined to use carbon markets. The electricity sector is expected to play a leading role in cutting emissions when nations put a price on carbon. This chapter describes the problem of global climate change and the pivotal role of the electric power industry. The description of climate change extends to the end of the century because of the long-lived effect of CO2 emissions. The description of the electricity sector concentrates on the next two decades, a period in which the electricity sector could lead the way in reducing emissions.

EBL

14.1. Introduction This chapter begins with a brief description of the global warming which has been attributed to the unprecedented accumulation of CO2 and other GHG in the atmosphere. This chapter concentrates on CO2 , the principal greenhouse gas. Results from major studies are summarized to reveal the likely impacts of a continued increase of atmospheric CO2 . Scientists and policymakers are calling for limits on the emissions, and this chapter summarizes the limits imposed by current policies, by proposed legislation, and in scientiﬁc studies. Some of the goals stretch out over the remainder of the century. They call for massive reductions in CO2 emissions if we are to stabilize atmospheric CO2 and reduce the risks of global warming. Reducing CO2 emissions will be a century-long challenge. (Adapting to a warmer world is also an important, century-long challenge. But this chapter focuses on mitigation rather than adaptation.) 499

Competitive Electricity Markets

500

This chapter describes the pivotal role to be played by the electric power industry in the coming years. Policymakers around the world recognize the importance of the electricity sector because of its large emissions, around one-third of the CO2 emissions in the United States, for example. The power sector is also recognized for its ﬂexibility to use different primary sources of energy to generate electricity. There is a long history of electricity regulation in some nations, and this sector is a convenient place to introduce carbon policy. The policy options include a combination of regulations, targets, and economic incentives. The main economic incentives are a carbon tax, a carbon market, or a combination of both. Many economists favor the carbon tax, but the political factors in Europe and the United States favor the adoption of a carbon market. There are several market proposals before the US Congress, and the European nations have launched the Emissions Trading Scheme (ETS), a 3-year mandatory market with trading among six industry sectors. This chapter describes the early ETS prices and some of the lessons to be drawn from the carbon trading. The chapter then describes legislative proposals for carbon markets in the United States. One of the more carefully studied proposals is Senate Bill 139 (S139), the Climate Stewardship Table 14.1. List of acronyms C CARBCCX CDM CO2 CO2 e CPUC CT EC EIA EPA EPPA ETS EU EUA GCM GDP GHG IPCC IGSM MIT NAP NASA NCEP NREL PEI RGGI S139 UK UNFCCC US WGA WECC

Carbon, the C in CO2 California Air Resources BoardChicago Climate Exchange Clean development mechanism Carbon dioxide; there are 3.67 tons of CO2 for each ton of C Weighted sum of GHG said to be equivalent to CO2 California Public Utilities Commission Combustion turbine European Commission Energy Information Administration (US) Environmental Protection Agency (US) Emissions Prediction and Policy Analysis Model (MIT) Emissions Trading Scheme (Europe) European Union European Union Allowance (carbon allowance in the ETS) General Circulation Model Gross domestic product Greenhouse gas, such as CO2 , methane, and nitrous oxide Intergovernmental Panel on Climate Change Integrated Global Systems Model Massachusetts Institute of Technology National Allocation Plan (Europe) National Atmospheric and Space Administration (US) National Commission on Energy Policy (US) National Renewable Energy Laboratory (US) Princeton Environmental Institute Regional Greenhouse Gas Initiative (northeastern US) Senate Bill 139, the Climate Stewardship Act of 2003 United Kingdom United Nations Framework Convention on Climate Change Unites States (of America) Western Governors Association (US) Western Electricity Coordinating Council (US and Canada)

EBL

Global Climate Change and the Electric Power Industry

501

Table 14.2. List of units aGW BTU ºC ºF E GT GW kw kwh MMTC MMTCO2 MTCO2 MT MW MWH ppmv $/MWH $/MTC $/MTCO2

Average GW, the energy from 1 GW operating for all hours in a year British Thermal Unit, a measure of energy Temperature in degrees centigrade Temperature in degrees Fahrenheit: a change of 1 C is the same as 18 F Euro (conversions in this chapter based on $1.27 = E1) Gigaton, a measure of weight = billion metric tons Gigawatt, a measure of capacity = 1000 MW Kilowatt, a measure of capacity = 1000 watts Kilowatt-hour, electric energy from 1 kw operating for 1 h Million metric tons of C Million metric tons of CO2 Metric ton of CO2 Metric Ton = 1000 kg Megawatt, a measure of capacity = 1000 kw Megawatt-hour, a measure of electric energy Parts per million by volume $ per MWH to measure electricity prices ($ are US $) $ per metric ton of C, used to show the cost of a carbon allowance $ per metric ton of CO2 ; divide by 3.67 to convert to $/MTC

Act of 2003. Acronyms are listed in Table 14.1; units of measure are listed in Table 14.2. The chapter describes simulated impacts of S139 for the US energy system and for the western electricity system. Compared to other sectors of the economy, the power sector is expected to lead the way in lowering CO2 emissions. Most of the reduction will come from a decline in coal-ﬁred generation. The industry would compensate for the loss of coal generation by improved efﬁciency of electricity use, increased generation from advanced technologies (not yet commercial), and increased generation from renewable technologies that do exist today. The study of the electric system in the western United States shows a major contribution from wind and biomass generators, technologies that exist today. The analysis of the western system indicates that the electricity sector could achieve dramatic reductions in CO2 emissions, and it could do so with existing technologies.

EBL

14.2. Global Warming The world is getting warmer, both in the atmosphere and in the oceans. The year 2005 was the warmest year in over a century, according to a recent NASA report (2006a). Data gathered from 1995 to 2006 have revealed that 11 of the past 12 years have ranked among the 12 warmest years in the instrumental record. Observations since 1961 have shown that the average ocean temperature has increased to depths of at least 3000 m. This warming causes the seawater to expand, thus contributing to rising sea levels (IPCC, 2007, p. 4). This section reviews the evidence of global warming, drawing heavily on the reports of the Intergovernmental Panel on Climate Change (IPCC). The IPCC was formed in 1988 and is described in Box 14.1. The clearest and most emphatic statement on global warming appeared in the IPCC’s most recent summary for policymakers released in February 2007: Warming of the climate system is unequivocal, as is now evident from the observations of increases in global average air and ocean temperatures, widespread melting of snow and ice, and rising global mean sea level. (IPCC, 2007, p. 4)

502

Competitive Electricity Markets

Box 14.1. The Intergovernmental Panel on Climate Change The IPCC was formed in 1988 by the World Meteorological Organization and the United Nations Environmental Program. The 2001 report was the third in a series of assessments of climate change. The summary of the fourth assessment was released on 2 February 2007. The assessments are considered to adhere to a high standard of objective reporting of the science of climate change. The reports must rely on peer-reviewed research, and the summaries must be unanimous, approved by all participating delegates, some of whom serve as representatives of their governments. Many view the IPCC as a scientiﬁc, consensus-building organization whose reports are conservative and cautious, thus avoiding the tendency to overstate the risks of climate change and the role of anthropogenic emissions. Some scientists believe that the reports understate the extent of climate change because of the modeling assumptions (Hansen et al., 2007; Kerr, 2007). Others believe the likely impacts are understated because of the difﬁculty in describing scientiﬁc ﬁndings while still achieving consensus from all delegates (Flannery, 2005). The IPCC process is slow and difﬁcult, and the reports may not represent the best and latest science. Nevertheless, the reports carry great weight with media and governments precisely because they present a consensus view.

EBL The IPCC 2001 report estimated that global average surface temperature has increased by about 0.6ºC during the twentieth century. The 2001 report noted that the decade of the 1990s was the warmest decade and that 1998 was warmest year of the previous millennium. The IPCC reported a decline in snow cover and a widespread retreat of mountain glaciers in non-polar regions during the same period. The panel used tide gauge data to learn that global average sea levels rose between 0.1 and 0.2 m during the century. And ﬁnally, the panel reported that global ocean heat content has increased since the late 1950s. The IPCC concluded that a combination of both natural and anthropogenic forces are responsible for the global warming. The natural causes include solar variations and volcanic activity, forces that are outside our control. The anthropogenic causes are subject to our control; they include the release of CO2 and other gasses which accumulate in the atmosphere. A major challenge for the scientists is judging the relative importance of the anthropogenic causes. The conclusion in 2001 was that anthropogenic emissions needed to be included for the scientists to explain the extent of global warming. This ﬁnding marked an important shift in scientiﬁc consensus from the report issued 6 years earlier. According to the IPCC’s 1995 assessment: “The balance of evidence suggests a discernible human inﬂuence on global climate.” A stronger conclusion was not possible in 1995 because the “anthropogenic signal was still emerging from the background of natural climate variability.” However, with more data, better attribution studies and improved climate modeling, the consensus in 2001 was that: Most of the observed warming over the last 50 years is likely to have been due to the increase in greenhouse gas concentrations. (IPCC, 2001, p. 10)

Global Climate Change and the Electric Power Industry

503

In the fourth assessment report, released in February 2007, the IPCC makes use of longer and improved records, improved simulation studies, and new attribution studies. The 2007 report is also notable for the consistent emphasis on the uncertainties that make measurements, simulation, and attribution difﬁcult. The Summary for Policymakers of the fourth assessment concluded that: Most of the observed increase in globally averaged temperatures since the mid-20th century is very likely due to observed increase in anthropogenic greenhouse gas concentrations. This is an advance since the Third Annual Assessment’s conclusion that “most of the observed warming over the last 50 years is likely to have been due to the increase in greenhouse gas concentrations.” Discernible human inﬂuences now extend to other aspects of climate, including ocean warming, continental-average temperatures, temperature extremes and wind patterns. (IPCC, 2007, p. 8) The remainder of this chapter draws heavily on the IPCC description of global warming. It also makes extensive use of reports from the MIT Joint Program on the Science and Policy of Climate Change. Additional information is available from a variety of reports and articles that appeared in 2006 (Bowen, 2006; NASA, 2006a, b; Stern Review, 2006; The Economist, 2006; Technology Review, 2006). Readers interested in the long history of scientiﬁc research on climate change are referred to The Discovery of Global Warming (Weart, 2003). Other popular books on the subject are authored by Schneider (1990), Flannery (2005), and Gore (2006). 14.3. The Greenhouse Effect

EBL

The atmosphere holds CO2 and other gasses which act to trap part of the infrared waves that reradiate from the surface of the earth. CO2 is the most prevalent GHG, but other gasses are important as well. They include methane and nitrous oxide whose concentrations are far above the pre-industrial levels. Anthropogenic methane emissions arise primarily in agriculture and partly from the energy sector. Anthropogenic nitrous oxide emissions come from vehicles and power plants, but mainly from agriculture. The layer of GHG acts like the glass in a greenhouse to capture a portion of the outgoing infrared radiation. As the concentrations of the gasses increase in the atmosphere, a greater fraction of the outgoing radiation is trapped, and the earth is warmed. This chapter concentrates on CO2 emissions, the principal GHG. Pre-industrial concentrations of CO2 were around 280 parts per million by volume (ppmv). The growth in atmospheric CO2 is well known because of the measurements begun in 1958 at Manua Loa on the big island of Hawaii. The ﬁrst measurements revealed concentrations around 315 ppmv. They also showed the seasonal cycle in atmospheric CO2 caused by the spring-time growth and the fall-time decay in biomass in the Northern Hemisphere. As the measurements continued, the upward trend quickly grew well beyond the seasonal variations. By 2005, the CO2 concentration had climbed to around 380 ppmv. A geologic perspective is provided by ice core data which allows scientists to infer the atmospheric CO2 concentrations and global temperatures over the past 400 000 years. Summaries of the ice core ﬁndings are provided by the United Nation’s Environmental Program (UNEP, 2006) and in special features published in The Economist (2006) and in Technology Review (2006). The Vostock ice core measurements show variations in atmospheric CO2 from around 180 to 300 ppmv. These large changes occurred during cycles that lasted around hundred thousand years and which have been attributed to minute

504

Competitive Electricity Markets

changes in the Earth’s orbit around the sun (Bowen, 2006). The measurements show that atmospheric CO2 , earth temperature, and sea levels rose and fell roughly in tandem during these cycles. The periods of low CO2 and temperature correspond to ice ages or “glacial” periods. The periods of high CO2 and high temperature correspond to the “interglacial” periods. The distinctive feature of the industrial era is the rapid rise in atmospheric CO2 to around 380 ppmv, a concentration far above any measurements for the past 400 000 years. The IPCC (2001) assessment reported that the concentration of CO2 in the year 2000 had not been exceeded “during the past 420,000 years and likely not during the past 20 million years.” The IPCC (2001) also reported that the rate of increase of atmospheric CO2 in the year 2000 was “unprecedented during at least the past 20,000 years.” The unprecedented accumulation of CO2 in the atmosphere presents long-lasting problems which will require serious action for the remainder of this century. To appreciate the century-long challenge, it is useful to review the ﬂows and accumulation of carbon that comprise the global carbon cycle. 14.4. The Global Carbon Cycle Figures 14.1 and 14.2 show the accumulation of CO2 in the atmosphere as part of the global carbon cycle. Figure 14.1 is a UNEP schematic showing the carbon ﬂows in a visual manner. Figure 14.2 summarizes the key stocks or storage of carbon in the system. The stocks in Fig. 14.2 are represented by the rectangles with estimates of current storage in GT, gigatons of carbon. The arrows depict the annual ﬂux or ﬂow of carbon between the atmosphere and the terrestrial system and between the atmosphere and the ocean system. These ﬂows are measured in GT/yr of carbon, and they are depicted in Fig. 14.2 with numbers rounded off to the nearest GT/yr.

EBL

Fig. 14.1. The global carbon cycle. Source: United Nations Environmental Program (UNEP), at: http://www.unep.org/. Similar depictions are available from NASA, at: http://earthobservatory. nasa.gov/Library/Carboncycle/carbon_cycle4.html, IPCC (2001), and from the Royal Society (2005) policy document on ocean acidiﬁcation.

Global Climate Change and the Electric Power Industry

505

Atmosphere 750 GT Anthropogenic emissions 6 GT/yr

Primary production 121 GT/yr Respiration 60 GT/yr

CO2 released from ocean 88 GT/yr

Ocean upper layer 600 GT

Biomass 600 GT Biomass litter decay 60 GT/yr

CO2 dissolves in ocean 90 GT/yr

Decomposition 60 GT/yr

Soils 1600 GT

Ocean middle layer 7,000 GT Ocean deep layer 30,000 GT

Fig. 14.2. Depiction of carbon ﬂows and storage from educational websites. Sources: These are approximate estimates obtained from a combination of educational sources listed in Fig. 14.1

The many ﬂows in Fig. 14.2 make it difﬁcult to anticipate the growth in atmospheric carbon over the coming decades. For example, how would one estimate the carbon in the atmosphere if anthropogenic emissions were to double over the next 50 years? It is also difﬁcult to estimate how carbon in the atmosphere would change if the emissions were reduced. One way to anticipate the changes in atmospheric carbon is to imagine that the only ﬂow to or from the atmosphere is the anthropogenic emissions, the 6 GT/yr highlighted in Fig. 14.2. The anthropogenic load is growing at around 1.4%/yr. With this trend, the load would reach 7.5 GT/yr in around 15 years. During this short interval, around 100 GT would be added to the atmosphere. If anthropogenic emissions were then to remain constant at 7.5 GT/yr over the next 100 years, another 750 GT would be added to the atmosphere. Atmospheric storage of CO2 would then be over twice as large. The 750 GT of CO2 in the atmosphere corresponds to a concentration of 352 ppmv. Were the atmospheric storage to double, the concentration would double as well. Atmospheric CO2 would be over 700 ppmv in just over 100 years. This simple example is a useful starting point to anticipate the CO2 concentrations that might occur in the next 100 years. The calculation is made simple by focusing on only one ﬂow in Fig. 14.2. It is as if the anthropogenic emissions that enter the atmosphere remain there forever. However, we know that CO2 emissions that enter the atmosphere are removed rather quickly by primary production of terrestrial biomass or absorption by the ocean’s upper layer. For example, Houghton (2004, p. 53) explains that a typical CO2 molecule is exchanged with the ocean in less than a year after entering the atmosphere. But the removal of CO2 from the atmosphere does not mean that the carbon is removed from the system. Rather, the carbon circulates through the system, re-entering the atmosphere at one point, exiting the atmosphere again at a later point, and returning to the atmosphere yet again further in the future. The overall effect of these circular ﬂows is that CO2 emitted into the atmosphere today “will contribute to increased concentration of this gas and the associated climate change for over a hundred years” (Houghton, 2004, p. 227). The dynamic consequences of these circular ﬂows are difﬁcult to anticipate. The difﬁculties complicate the analysis of scientists as well as the thinking of policymakers and the general public. The challenge of global warming makes it imperative that both the general

EBL

506

Competitive Electricity Markets

public and policymakers develop an intuitive understanding of CO2 accumulation in the atmosphere.1 If policymakers and the public lack an intuitive base, they are likely to have difﬁculties in understanding and acting upon the latest scientiﬁc ﬁndings. Our intuition can be improved by concentrating on the net ﬂows from the atmosphere to the terrestrial system and the ocean system. The left side of Fig. 14.2 shows the net ﬂows to the terrestrial system. Total ﬂows out of the atmosphere exceed the inﬂow by 1 GT/yr. This imbalance suggests that around 1 GT/yr of carbon is added to the stock of biomass or soil. Figure 14.2 indicates that the carbon stored in the biomass would grow over time (perhaps due to extensive reforestation of previously cleared land). The right side of Fig. 14.2 shows the ﬂows from the atmosphere to the ocean. The ﬂow out of the atmosphere exceeds the inﬂow by 2 GT/yr. The total net ﬂow out of the atmosphere is 3 GT/yr, which means that natural processes are acting to negate approximately half of the current anthropogenic load (IPCC, 2001; Socolow et al., 2004). Figure 14.2 provides a starting point for anticipating future accumulation in the atmosphere. As the use of fossil fuels grows over time, the anthropogenic load will increase. But scientists do not think that natural processes can continue to negate 50% of an everincreasing anthropogenic load. On the terrestrial side of the system, there are limits on the net ﬂow associated with reforestation of previously cleared land (Socolow et al., 2004). And there are limits to the carbon sequestration in plants and soils due to carbon–nitrogen constraints (Gill et al., 2006). On the ocean side of the system, the current absorption of 2 GT/yr is already sufﬁciently high to disrupt the chemistry of the ocean’s upper layer (Sarmiento et al., 1995; Socolow et al., 2004; The Royal Society, 2005). The extra 2 GT/yr contributes to problems of acidiﬁcation, and it changes the balance of dissolved minerals. Higher CO2 can reduce the concentration of carbonate, the ocean’s main buffering agent, thus affecting the ocean’s ability to absorb CO2 over long time periods.

EBL

14.5. Climate Models and Uncertainty Scientists use a variety of models to keep track of the greenhouse gasses and their impact on the climate, as explained in Box 14.2. Some of the models combine simulations of the atmosphere, soils, biomass, and ocean response to anthropogenic emissions. The more developed models include CO2 , methane, nitrous oxides, and other GHG emissions, and keep track of their changing concentrations in the atmosphere. The well-developed models also keep track of sulfates (short-lived pollutants which arise primarily from the release of sulfur dioxide from power plants). The sulfates act to reﬂect sunlight back into space, thereby contributing to the cooling of the planet. Inclusion of the sulfates allowed scientists to explain the “curious enigma” of the 1950s–1970s. This was a period of rising CO2 in the atmosphere, but the temperature data showed a cooling trend. Climate models helped scientists understand this enigmatic situation when they clariﬁed the role of aerosols in masking the temperature increases that would normally be expected from the rising CO2 concentration. 1 Building our intuition is not easy, as revealed by careful experiments with highly educated adults reported by Sterman and Sweeney (2007). When asked about CO2 accumulation in the atmosphere, the subjects tended to misperceive the sluggish response of atmospheric CO2 . The experiments showed that many subjects tend to match the pattern of atmospheric CO2 with the pattern of anthropogenic emissions. As an example, a “pattern-matching” subject would conclude that atmospheric CO2 will decline this year if anthropogenic emissions were to decline this year. This type of thinking can lead to a “wait and see” attitude about climate policy that is inappropriate when today’s emissions contribute to higher concentrations over the next hundred years.

Global Climate Change and the Electric Power Industry

507

Box 14.2. Models of the climate system A wide variety of models are used to improve our understanding of climate change. All the models provide a useful perspective on the highly non-linear dynamics of the climate system. Claussen et al. (2000) classiﬁes the models according to the degree of complexity: simple, intermediate, and comprehensive. The simple models represent the physical concepts in a tutorial fashion. They are sometimes called “box models” since they represent the storage in the system by highly aggregated stocks like those shown in Fig. 14.2. The parameters are usually selected to match the results from more complicated models. The simple models can be simulated faster on the computer, and the results are easier to interpret. This makes them valuable in conducting extensive sensitivity studies and in scenario analysis. A primer on climate modeling and the value of simple models is provided by the IPCC (1997). The comprehensive models are maintained by large research centers, including NASA, NCAR, NOAA, and the Hadley Center in the United Kingdom. The term “comprehensive” refers to the goal of capturing all the important processes and simulating them in a highly detailed manner. The models are sometimes called General Circulation Models (GCMs). They can be used to describe circulation in the atmosphere or the ocean. Some models simulate both the ocean and atmospheric circulation in a simultaneous, interacting fashion. They are said to be “coupled general circulation models” (CGCMs) and are considered to be the “most comprehensive” of the models available (Claussen et al., 2000). They are particularly useful when a high spatial resolution is required. However, a disadvantage of the CGCMs is that only a limited number of multi-decadal experiments can be performed even when using the most powerful computers. Intermediate models help scientists bridge the gap between the simple and the comprehensive models. Claussen et al. (2000) describe 11 models of intermediate complexity. These models aim to “preserve the geographic integrity of the Earth system” while still providing the opportunity for multiple simulations to “explore the parameter space with some completeness. Thus, they are more suitable for assessing uncertainty.” This chapter uses the intermediate model developed at the Massachusetts Institute of Technology (MIT).

EBL

This chapter draws on the results of the Integrated Global Systems Model (IGSM) developed by MIT (MIT, 2006; Prinn et al., 1999; Reilly et al., 1999). The IGSM analysis of the uncertainty associated with future climate change (Webster et al., 2003) is particularly useful. The analysis began with an estimate of anthropogenic emissions growing to around 19 GT/yr by 2100. The focus of the analysis was the band of uncertainty in future conditions, both the anthropogenic emissions and the climatic effects of those emissions. The mean projection of atmospheric CO2 was around 700 ppmv by 2100. Figures 14.3 and 14.4 summarize the MIT analysis. Figure 14.3 shows the mean result over the century; Figure 14.4 displays the uncertainty in the CO2 concentration at the end of the century. Figure 14.3 puts the mean result of the MIT study in perspective by showing three different ways to accumulate anthropogenic emissions. The top curve represents the simplest

Competitive Electricity Markets

508

Atmosphere CO2 (ppmv) 900 850 800 750 700 650 600 550 500 450 400 350 300 250 2000

2050

2100

Accumulation of all emissions Accumulation with 3 GT/yr negated

EBL

Accumulation with up to 4 GT/yr negated Fig. 14.3. Accumulations of CO2 in atmosphere with anthropogenic loads growing from 6 to 19 GT/yr during the century. Sources: Author’s calculations to interpret the aggregate results published by Webster et al. (2003).

possible situation. It shows the growth in atmospheric CO2 with the assumption that there is no net removal of CO2 from the atmosphere by natural processes. Under this simplifying assumption, atmospheric CO2 would grow to 880 ppmv by the year 2100. The middle curve accumulates the anthropogenic emissions with the assumption that the natural processes continue to negate 3 GT/yr of the anthropogenic load: atmospheric CO2 would grow to 740 ppmv. The lowest curve provides the best match with mean projection from MIT’s IGSM. It was generated with the assumption that net ﬂows out of the atmosphere increase from 3 to 4 GT/yr during the ﬁrst half of the century and remain at 4 GT/yr for the second half of the century. This assumption allows the model to come quite close to the IGSM result: CO2 concentration grows to approximately 700 ppmv by the end of the century. This close match helps one interpret the results of the IGSM calculations. The results make sense if the natural processes are currently at about 75% of their satiation limit. As more CO2 accumulates in the system, net removal will reach a limit of around 4 GT/yr, and the removal would be split evenly between the terrestrial and ocean systems (Webster et al., 2003, p. 310). Figure 14.3 shows atmospheric CO2 growing from 350 to 700 ppmv in a century. The MIT IGSM mean estimate of the temperature impact is a 24 C warming (relative to the temperature in the year 1990). The 24 C of warming is a major impact (i.e. four times higher than the 06 C observed in the previous century). The 24 C of warming is within the range of estimates by Jim Hansen, Director of NASA’s Goddard Institute for Space

Global Climate Change and the Electric Power Industry

509

Range (5% to 95%) of estimated temperature impacts by 2100 6.0

5.0

4.0

3.0

2.0

1.0

0.0 250

350

450

550

650

750

850

950

1,050

1,150

Range (5% to 95%) of estimated CO2 concentrations in 2100

EBL

Business as usual

Stabilization policy

Fig. 14.4. Summary of the 5, 50, and 95% range of impacts from Webster et al. (2003). (The temperature impact is relative to the temperature in 1990.)

Studies (Bowen, 2006). He predicted that a continued growth of greenhouse gas emissions at current rates would cause global temperature to increase by 2 − 3 C in this century. Some readers will be familiar with “equilibrium climate sensitivity,” the IPCC term for the temperature impact from a doubling of CO2 concentration. This closely related concept is explained in Box 14.3. The main purpose of the MIT analysis was to show the uncertainty in the climate impacts. Webster et al. (2003) presented statistical analysis of 250 simulations with uncertainty in the model parameters. (Both the anthropogenic emissions and the impact of those emissions were subject to great uncertainty.) Figure 14.4 summarizes the statistical analysis by showing the 5, 50, and 95% estimates of CO2 concentrations and temperature impacts by the end of the century. The “business-as-usual results” assumes no policy intervention to limit emissions. The most likely result is a CO2 concentration around 700 ppmv and a global average surface temperature that is 24 C above the temperature in 1990. However, with a policy to control emissions, there would be a cap on emissions with the goal of stabilizing atmospheric CO2 at around 550 ppmv. The most likely result is now a CO2 concentration of 512 ppmv and a global average surface temperature that is 1.7ºC above the temperature in 1990. Figure 14.4 shows that the 5–95% range is greatly reduced with the stabilization policy. The atmospheric CO2 could range from 466 to 580 ppmv, and the temperature impact would range from 0.8 to 32 C.

510

Competitive Electricity Markets

Box 14.3. Equilibrium climate sensitivity The temperature impact from a doubling of CO2 concentration is mentioned frequently in IPCC reports. The scientists use the term “equilibrium climate sensitivity” to describe the temperature increase to be expected from a doubling of CO2 concentration relative to the pre-industrial concentration of 280 ppmv. (The impact assumes that the system remains in equilibrium at 560 ppmv.) The IPCC 1990 assessment gave a best estimate of climate sensitivity at 25 C (Houghton, 2004, p. 120). The best estimate in their most recent assessment (IPCC, 2007, p. 9) was 3 C with a range from 2 to 45 C. This measure of sensitivity is useful in comparing different models and in comparing the different assessment reports. However, the term is not entirely descriptive of the doubling of CO2 concentration in Fig. 14.3 since the simulation begins the century at 350 ppmv, and the concentration is still growing at the end of the century.

The 5–95% uncertainty bands portray the large range of impacts from the uncertain parameters in the MIT model. But parametric uncertainty is only part of the uncertainty. Changes in the fundamental structure of the model could also change the bands of uncertainty. Examples include the addition of new pollutants (i.e., such as the aerosols mentioned previously) or the inclusion of new feedback effects between the atmospheric and terrestrial systems. These sources of uncertainty are much more difﬁcult to quantify. Ultimately, statements about structural uncertainty come down to the scientists’ intuition on whether the omitted structure will act in a stabilizing or a destabilizing manner. In some cases, adding new relationships to a model will close negative feedback loops, which can act to stabilize the simulated system (Ford, 1999; Luoma, 1991). In many of these cases, the new structure could lead to a narrower band of uncertainty. However, the customary process of model development is to ﬁrst include most of the pervasive, well-understood, negative feedback loops at work in the system. The less-understood feedback loops are often left to future work (when more evidence about their role becomes available). These omitted feedbacks can often be positive feedback loops that act to destabilize the system, as explained in Box 14.4. Rapid climate change is now evident in the past record and is taken as a serious possibility in the future. Weart (2007) cautions that such changes are not necessarily explainable with the main GCMs. Drawing on a National Academy study (NAS, 2002), he cautions that “The abrupt changes of the past are not fully explained yet and climate models typically underestimate the size, speed and extent of those changes. Hence, climate surprises are to be expected.” Positive feedback effects are difﬁcult to simulate, but their destabilizing effects are important to consider when scientists and policymakers think about the uncertainty of the system. This is why scientists often conclude their quantitative analysis of uncertainty with a qualitative assessment of the possible surprises from relationships that have not yet been simulated in a model. An example is the concluding remarks

EBL

Global Climate Change and the Electric Power Industry

511

Box 14.4. Destabilizing feedback and rapid climate change A closed chain of cause and effect that acts to destabilize a system is sometimes called a positive feedback loop. (The term “positive” comes from control theory. It does not denote that the feedback will lead to changes that are good or bad.) Understanding the role of positive feedback has been crucial to scientists’ research and eventual “discovery” of rapid climate change. Weart (2003, 2007) explains that “swings in temperature that were believed in the 1950s to take tens of thousands of years, and in the 1980s to take hundreds of years, were now found to take only decades.” Examples of positive feedback loops include: •

•

•

•

•

Methane from permafrost: Higher temperatures can cause the permafrost to shrink, releasing the methane embedded in the clathrate sediments to the atmosphere. More methane in the atmosphere could lead to further warming and still greater shrinking of the permafrost. Methane from bogs and swamps: Higher temperatures can accelerate the decomposition of dead organic matter in bogs and swamps, also releasing methane to the atmosphere. Water vapor: Higher temperatures lead to more water vapor in the atmosphere which can lead to an increase in long wave absorption. With more absorption, there could be still greater warming and more water vapor in the air. Soil decomposition: Higher temperatures tend to cause faster decomposition of soil carbon, releasing more CO2 into the atmosphere, thus trapping more radiation and increasing the temperature still further. Sea ice/albedo ﬂip: The sea ice has a higher albedo than the surrounding water. As the ice melts, there is an increase in ice-free water which leads to more heat absorbtion. This increases the polar temperatures causing still further melting of the sea ice.

EBL

The water vapor and soil decomposition feedbacks involve a combination of stabilizing and destabilizing feedbacks acting in tandem. With increased water vapor, for example, there may be greater short wave reﬂection which acts as a stabilizing feedback effect. The relative strength of the water vapor feedback effects is said to be a key factor inﬂuencing climate sensitivity (IPCC, 2007, p. 9). The soil decomposition also involves a combination of destabilizing and stabilizing effects. Higher CO2 concentrations can lead to greater biomass growth (due to the “fertilization effect,” subject to sufﬁcient nitrogen in the soil to support the growth). This is a stabilizing feedback since increased biomass growth removes CO2 from the atmosphere. Sorting out the relative power of the soil carbon feedbacks requires detailed analysis with “fully coupled” models (Cox et al., 2000; Govindasamy et al., 2005; Jones et al., 2003; Kump, 2002). Such analyses show the possibility for soil carbon to change from a net sink to a net source of carbon to the atmosphere. The possible reversal of the net ﬂow from the atmosphere to the soils is a major source of uncertainty in the system, one that is not easily resolved without further research (Govindasamy et al., 2005; Kump, 2002). The destabilizing effect of the sea ice/albedo ﬂip is described in detail by Hansen et al. (2007) and discussed in the news focus editorial by Kerr (2007). Hansen

512

Competitive Electricity Markets

emphasizes the loss of sea ice to make sense of the rapid temperature increases at the terminations of the last two ice ages. He argues that the IPCC models do not represent the amplifying effect of the sea ice feedback and their projections understate the possibility of a rapid rise in sea level

about the analysis of parametric uncertainty published by Webster et al. (2003, p. 317): As with all investigations of complex and only partially understood systems, the results presented here must be treated with appropriate caution. Current knowledge of the stability of the great ice sheets, stability of thermohaline circulation, ecosystem transition dynamics is limited. Therefore abrupt-changes or “surprises” not currently evident from model studies, including our uncertainty studies summarized here, may occur. 14.6. Dealing with Uncertainty Some may use uncertainty as an excuse to ignore global warming. They might focus on the low end of the range of impacts and argue that 10 C warming is not worth the effort to curb emissions. Or they might interpret the wide range of uncertainty as a sign that the scientists’ predictions are too uncertain to justify immediate action. They might argue that it makes better sense to postpone difﬁcult decisions about emissions reductions until the predictions can be made with greater certainty. Postponing difﬁcult decisions is one option for dealing with a highly uncertain system. Indeed, in some situations, a “wait and see” approach can be the best strategy for risk management. The problem with a “wait and see” strategy on global warming is that investment decisions are being made today that can lead to increased fossil fuel consumption over the operating lives of the investments. The emissions from the new faculties will enter the atmosphere, circulate through the complex system shown in Figs 14.1 and 14.2, and their impact will be felt over the remainder of the century. With such unpredictable, long-lived effects, a different approach to risk management is needed. It makes more sense to confront the uncertainty in anticipating future climate impacts. It is particularly important to limit emissions that could lead to the high concentrations of CO2 and trigger some of the destabilizing feedbacks. Indeed, this danger has been formally recognized by 190 nations which have signed the United Nations Framework Convention on Climate Change (UNFCCC). It aims to stabilize GHG concentrations at a level that would prevent “dangerous anthropogenic interference” with the climate system (UNFCCC, 1992; Mastrandrea and Schneider, 2004). The statistical analysis of the IGSM provides useful insight on the range of impacts that could be expected in either a business-as-usual scenario or a stabilization scenario. Equally useful information on uncertainties is provided in the assessments by the IPCC. Much of the third assessment is devoted to a better understanding of the wide bands of uncertainty. The report shows ranges of estimates for the historical record when looking back in time, and it is careful in summarizing expectations for important, extreme events by selecting words that convey their expectation while reminding policymakers of the uncertainty. For example, the third assessment reports that it is “very likely” that higher maximum temperatures and more hot days will occur over nearly all land areas in the twenty-ﬁrst

EBL

Global Climate Change and the Electric Power Industry

513

century. The fourth assessment continues to provide a wide band of projections into the future, and it takes the same care in describing extreme weather events. In the case of hotter days, the 2007 report notes that it is “virtually certain” that higher maximum temperatures and more hot days will occur over most land areas during the twenty-ﬁrst century. The IPCC’s views into the future are based on a collection of scenarios which allow policymakers to anticipate projected impacts across different “storylines” involving growth, development, technology, and sustainability. The wide range of conditions allow for a wide range of simulated outcomes. The IPCC uses this approach to emphasize the uncertainty in future outcomes, not to single out a particular scenario. Rather, they advise that all scenarios “should be considered equally sound” (IPCC, 2001, p 18; 2007, p. 14). The analysis by Webster et al. (2003) took a different approach. The uncertainties in future projections were based on statistical analysis of many simulations with variations in the values assigned to the input parameters. When facing a future without limitations on carbon emissions, the statistical analysis showed that atmospheric CO2 could climb to 1032 ppmv. The high-end temperature impact would be 49 C. With emissions controls, however, the high-end impacts are greatly reduced. Instead of atmospheric concentrations over 1000 ppmv, the concentrations are around 580 ppmv. Instead of 49 C warming, the high-end warming would be 32 C. Further analysis of the narrowing of uncertainty through control of emissions is provided by Forest et al. (2004). This reduction in uncertainty is achieved by policy controls on the anthropogenic emissions. The policy scenario assumes that emissions are controlled in such a manner as to stabilize atmospheric CO2 at around 550 ppmv. The risk of global warming is greatly reduced but not eliminated. The reduction in the high-end impacts is one of the main reasons why scientists and policymakers are calling for the imposition of limits on anthropogenic emissions. An example is the Stern Review on the Economics of Climate Change which has been the subject of much debate since its release in 2006. Of particular interest was the estimated cost of living in a warmer world if no action were taken to curb emissions. The Stern Review estimated the impacts of global warming in the range of 5–10% of the world’s gross output in a business-as-usual scenario. They estimated the costs of reducing emissions at around 1% of the world’s gross output. The cost of inaction would be ﬁve to ten times higher than the cost of cutting emissions. This ﬁnding has sparked much debate, and there are many who disagree with Stern’s assumptions, particularly the assumptions about impacts far in the future and the appropriate way to discount such impacts when making monetary comparisons. Readers interested in criticisms of the Stern Review are referred to Lomborg (2006), Mendelsohn (2006), Tol (2006), and Varian (2006). An excellent interpretation of the study and its implications for US policy is provided by Jacoby (2007). The key ﬁndings are summarized in Box 14.5.

EBL

Box 1 4.5. The Stern Review on the economics of climate change This report was issued in October 2006 in the United Kingdom (UK) by Sir Nicholas Stern, former chief economist at the World Bank. The report brings together a wide range of estimated impacts for different values of atmospheric GHG and draws attention to the wide range of uncertainties associated with the impacts. The Review

514

Competitive Electricity Markets

describes an aggregate measure of GHG concentrations in CO2 e (a weighted sum of greenhouse gasses equivalent to atmospheric CO2 .) With this measure, the current GHG concentration is around 430 ppmv CO2 e (compared to 380 ppmv for CO2 alone). The report predicted that, even with anthropogenic emissions of GHG limited to their 2006 value, atmospheric CO2 e would grow to 550 ppmv by the year 2050, and would continue to grow thereafter. In this scenario, average global temperatures would be 30 C higher than the pre-industrial level; crop yields would fall in many countries; hurricane intensities would increase; sea levels would rise; and the risks of rapid climate change could include the collapse of the West Antarctic ice sheet. The Review predicts even higher concentrations from growing emissions in a businessas-usual scenario, a scenario with a 40 − 50 C increase in global temperature. The Stern Review looked at the economic impacts of these changes. It noted that previous studies estimated economic impacts at around 0–3% permanent loss of global world output with 2 − 3 C warming. The report warned that the economic calculations must deal with the sobering possibility of 4−5 C warming and the effect of the rapid, irreversible impacts. The economic impacts are not easily estimated, given the unprecedented nature of the expected impacts, but the Stern Review made their best attempt since the goal was to provide a comparison of the costs of inaction with the cost of curbing emissions. Their best estimate of the simulated impacts was an average, permanent loss of 5–10% of global output. The report cautioned that the impact would not be evenly distributed around the world, and that poorer countries would experience permanent losses greater than 10%. It further cautioned that the estimated loss of output did not take into account non-market impacts such as damage to the environment and human health. It also failed to represent the existence of amplifying feedbacks such as the release of methane and the weakening of carbon sinks. It argued that putting these additional factors together would increase the total cost of the business-as-usual scenario to the equivalent of a 20% reduction in per capita consumption. The Review presented a variety of emissions paths that would allow atmospheric GHG to stabilize at 550 ppmv of CO2 e. They estimated that achieving these cuts in emissions would cost around 1% of gross world output by the year 2050, a cost which was described as “signiﬁcant but manageable.”

EBL

14.7. Goals for Emission Reductions Figure 14.5 summarizes some of the targets for emission reductions that have been adopted or proposed around the world. In many cases, the targets are speciﬁed relative to a country’s emissions in the year 1990. So, for ease of comparison, Fig. 14.5 uses 100 to denote emissions in the year 1990. Emissions have been growing at around 1.4%/yr. The upward curve shows the future emissions if this trend continues: emissions would reach 200 by 2040 and 400 by 2090. Figure 14.5 shows the great differences in the stringency of the targets. Some call for holding emissions constant; others call for dramatic reductions over time. Some targets apply to the next two decades; many extend to the year 2050; and some extend to the year 2100. However, when compared to the upward trend, all targets require major reductions in emissions.

Global Climate Change and the Electric Power Industry

515

MIT long term stabilization Path

250

Senate bill 139

225 Kyoto (2008 to 2012)

200 Stern review stabilization targets

175

California AB 32

150

Princeton Wedges Study: hold constant

125

Emissions if growth continues at 1.4%/yr

100

UK 60% reduction by 2050

75

Congress: 2008 emissions Congress: 50% reduction

50

Congress: 80% reduction

25 EU 20% reduction by 2020

0 1990 2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100

Fig. 14.5. Comparison of goals for emission reductions. (The 100 on the vertical axis represents emissions in the year 1990.)

EBL

The targets from the Kyoto Treaty are probably the best known. The treaty became effective in February 2005 and called for the Annex I countries to reduce emissions, on average, by 5% below 1990 emissions by the year 2008 and to maintain this limit through 2012. The Kyoto Protocol allows for a range of emission reductions (i.e., some countries will aim for reductions greater than 5%) and allows countries to purchase reductions achieved elsewhere (i.e., through a clean development mechanism, CDM). The extension of the Kyoto Protocol beyond 2012 is the subject of ongoing discussions. The Princeton study is also widely known, so it serves as a useful comparison with the many other goals. This particular target is shown in more detail in Fig. 14.6, and it is shown again in Fig. 14.7 to allow for a clear comparison with some of the more ambitious targets. The Princeton team compared a business-as-usual scenario for growing emissions with emissions that would allow atmospheric CO2 to stabilize at around 500 ppmv. This requires annual emissions to be held constant at 7 GTC beginning in 2004. The Princeton study described the difference between the constant emissions and the growth trend as a “stabilization triangle” comprising seven “wedges.” Each wedge depicts carbon reductions of 1 GTC per year by 2054. The study described the changes that could deliver the needed reductions. For example, improvement in building efﬁciency was estimated to deliver one wedge, and halting deforestation was estimated to deliver a half-wedge. Figure 14.7 puts the Princeton target in perspective. Their intent was to generate discussion about what could be achieved with current technologies. Since that time, there have been important publications about climate change and some political change in the US Congress. The Congress is currently considering legislation that would aim to reduce emissions far beyond the goal set in the Princeton study. Figure 14.7 illustrates with goals set over the time period from 2012 to 2050. The downward paths are simple, linear approximations used in a MIT study (to be described shortly). These goals were selected to be representative of several proposals currently under discussion in the US Congress.

Competitive Electricity Markets

Fossil fuel emissions (GtC/y)

516

16

B

14 12 10

Stabilization triangle

8 6

Continued fossil fuel emissions

4 2 0 2000

2010

2020

2030

2040

2050

2060

Year

Fig. 14.6. The stabilization triangle and seven wedges in the Princeton study. Source: Socolow et al. (2004).

250.0

Princeton wedges study

225.0

EBL

200.0 175.0

Emissions if growth continues at 1.4%/yr

150.0 125.0

Congress: 50% reduction

100.0 75.0

Congress: 80% reduction

50.0 25.0 0.0 1990

2000

2010

2020

2030

2040

2050

Fig. 14.7. Comparison of the goal in the Princeton study with the more aggressive goals representative of legislation under discussion in the US Congress.

Returning to Fig. 14.5, the solid line from 2010 to 2050 represents the stabilization path used in the MIT study described previously (Webster et al., 2003). The limit on emissions was imposed in modeling calculations designed to stabilize atmospheric CO2 at 550 ppmv or lower. The scenario assumed that the Kyoto emission caps are adopted by all countries by 2010. The policy assumed that the caps would be extended and then further lowered by 5% every 15 years. By the end of the century, the emissions would be 35% below the value in 1990.

Global Climate Change and the Electric Power Industry

517

Figure 14.5 shows the year 2100 target in the MIT study matches the corresponding target from the Stern Review. The Stern Review shows a variety of emission paths associated with stabilizing atmospheric GHG at 550 ppmv of CO2 e. The 2050 and 2100 goals are my selections to be representative of the various paths. The Stern Review was produced in the United Kingdom with the goal for worldwide reduction in emissions. Another UK report was issued a few years earlier. The Energy White Paper, presented to Parliament by the Secretary of State for Trade and Industry in February 2003, called on the United Kingdom to “put itself on a path to a reduction in CO2 emissions of some 60% from current levels by about 2050.” More recently, the 27 members of the European Union agreed to reduce CO2 emissions by 20% below 1990 levels by the year 2020.2 This chapter concentrates on Senate Bill 139, The Climate Stewardship Act of 2003. Figure 14.5 shows the S139 targets over the interval from 2010 to 2016. The bill called for an initial cap on emissions from 2010 to 2016. The cap would be reduced to a more challenging level in 2016, when the goal was to limit emissions to no more than the emissions from 1990. Figure 14.5 shows the S139 cap extending to the year 2025, the time period for studies described later in the chapter. There are several new proposals under consideration in the US Congress (NRDC, 2006). Some apply only to the electricity sector, but most apply to multiple sectors. They all call for cap-and-trade in carbon allowances, but the proposals differ widely in the stringency of the cap. Figure 14.5 shows three targets in the year 2050. These were used in a MIT (2007b) study to span the range of proposals under consideration: CO2 emission in 2050 could be at the value in 2008, 50% below the value in1990, or 80% below the value in 1990. Preliminary results of the MIT study are explained in Box 14.6.

EBL

Box 14.6. Analysis of proposals before the US Congress As of April 2007, there were six cap-and-trade proposals in various stages of consideration in the US Congress. For example, Senators McCain and Lieberman have introduced a new legislation, which is one of the four bills in the Senate. Two more proposals are under consideration in the House of Representatives. With so many proposals, the MIT team deﬁned three general scenarios to span the range of proposals. Each scenario assumes that the proposed market would open in the year 2012, and the economic analysis was conducted to the year 2050. (Climate impacts were then calculated to the year 2100 with the assumption that the emissions cap in effect in 2050 would remain in effect to the year 2100). The three scenarios were deﬁned as: •

Least ambitious: Emissions capped at the level in the year 2008. Total allowances issued over the interval from 2012 to 2050 would be 287 GTCO2 . • 50% Reduction: The initial cap would be at the 2012 emissions. The cap would then decline in a linear fashion to 50% below 1990 emissions by the year 2050. Total allowances issued over the interval from 2012 to 2050 would be 203 GTCO2 . 2

The 20% goal was announced at the EU Summit in Brussels in March 2007.

Competitive Electricity Markets

518

•

80% Reduction: The initial cap would be at the 2012 emissions. The cap would then decline in a linear fashion to 80% below 1990 emissions by the year 2050. Total allowances issued over the interval from 2012 to 2050 would be 167 GTCO2 .

Most of the legislative proposals are positioned between the 50% and 80% reduction scenarios. For this chapter, it is useful to share some of the preliminary results from the 50% reduction scenario. The MIT study predicts carbon allowance prices over the interval from 2015 to 2050 based on a core set of assumptions which include the option to bank allowances for future use. By 2020, carbon allowance prices are expected to reach 50$/MTCO2 ; by 2025, the price would be 61$/MTCO2 . These prices are noted here because they are almost exactly the prices mentioned later in the chapter for S139. The MIT report noted that the prediction of carbon allowance prices could also be interpreted as an estimate of the carbon tax needed to generate the same reduction in emissions. A carbon tax or an auction of carbon allowances would generate the same revenue. For example, a 50$/MTCO2 tax would generate $324 billion in revenues in the year 2020. This would amount to 14% of non-CO2 federal tax revenue. Or, if the revenues were used for disbursements, it would amount to tax disbursements of $3870 per household (assuming a family size of four). The MIT study concluded with a discussion of the importance of a global agreement on reducing CO2 emissions. If at all a cooperative solution is possible, a major strategic consideration in setting emissions targets in the United States should be their value in leading other major countries to take on similar efforts.

EBL

The ﬁnal target in Fig. 14.5 is placed in the year 2020 to represent the recent passage of Assembly Bill 32 in California. It calls for California’s emissions to be reduced to 1990 values by the year 2020. California is not the ﬁrst state to take action on carbon policy, but it is the ﬁrst to implement a policy that applies across major sectors of the state’s economy. Other states have taken important actions, but their ﬁrst steps concentrate on the electric power sector, as explained in Box 14.7.

Box 14.7. State and regional initiatives to control carbon emissions in the United States Several regions have embarked on carbon policy initiatives in the absence of national legislation. The states would create a variety of targets, regulations, and incentives. The ultimate targets apply to GHG emissions, but there are intermediate targets as well. They may call for a certain fraction of electricity generation to be supplied by renewable generators, for example. Most of these initiatives focus on the electricity sector. This makes sense given the states’ history of regulatory authority in electricity ratemaking and in power plant siting. One of the ﬁrst initiatives is in the Northeast/Mid-Atlantic states that have combined to form the Regional Greenhouse Gas Initiative (RGGI). It calls for cap-and-trade in fossil fuel-ﬁred generators starting in 2009. The initial cap would

Global Climate Change and the Electric Power Industry

519

limit electricity sector emissions to the value from 1990. The limit would then be reduced by 10% by the year until 2018. The RGGI goals are estimated to be achieved with negligible (and even beneﬁcial) impacts on the retail ratepayers (Ford, 2006; RGGI, 2006). The West Coast Governors Global Warming Initiative has issued various recommendations. They call for changes in standards and “smart growth” initiatives to increase use of rapid transit to reduce vehicle miles of travel. And they call for careful consideration to the development of a regional market-based carbon allowance program. The impact of their package of proposals is described by Ford (2006) and the Tellus Institute (2004). They involve a complicated combination of regulations, targets, and incentives, including consideration of cap-and-trade in allowances with a ﬁxed allowance price of 20$/MTCO2 . These proposals involved a combination of Washington, Oregon, and California. Since that time, Arizona and New Mexico have joined to declare a western regional climate action initiative (WRCAI, 2007). Oregon and Washington are both implementing climate change policies. The Oregon Strategy for Greenhouse Gas Reductions estimates that the state’s GHG would climb to 60% higher than 1990 emissions by the year 2025. Their goal is to arrest the growth in GHG by 2010 and to achieve reductions toward a benchmark for CO2 not exceeding 1990 levels. By 2020, Oregon aims to achieve a 10% reduction below 1990 GHG levels. The Governor of Washington just signed a new energy bill in May 2007. It calls for GHG emissions to be reduced by 50% relative to 1990 emissions during the next 40 years. It would also prevent Washington utilities from entering into long-term contracts with coal-ﬁred power plants that produce GHG emissions. California has announced two important initiatives in 2006. In February, the California Public Utilities Commission (CPUC) announced that it would develop a cap on GHG emissions for the state’s Investor Owned Utilities (IOUs) and nonutility companies. The CPUC (2007) explains that it will create a “load-based cap that encompasses all of the GHG emissions produced in the course of generating electricity to serve utility customers. Imported energy and power produced within California will be treated equally under this system.” The CPUC acknowledges the pioneering effort of the RGGI, but the California proposal would be organized as a “load-based” cap in which the cost of allowances are borne by the regulated IOUs which distribute electricity to retail customers. The CPUC has also proposed rules which would prohibit the IOUs from purchasing out-of-state generation that exceeds a carbon emissions standard. (This rule is viewed as a “stop-gap measure” to prevent a rash of coal plant development before the CPUC adopts limits on emissions.) This is an extremely important ruling since California has asserted control over carbon content of electricity delivered to serve the large loads in California. The other important California initiative is Assembly Bill 32, the Global Warming Solutions Act of 2006. AB32 sets the goal to reduce California’s GHG emissions to 1990 levels by the year 2020. (This is estimated to require a 25% reduction from emissions expected without the legislation.) AB32 would apply to all emissions, not just the emissions from the electricity sector. AB32 sets the goal, but it leaves many important decisions to be made later by either the state Environmental Protection Agency or the state California Air Resources Board (CARB). Detailed recommendations are provided by the market advisory committee of the CARB (2007).

EBL

520

Competitive Electricity Markets

14.8. Policies to Reduce Emissions Policymakers are considering a variety of targets, regulations and incentives to reduce GHG emissions. Targets for CO2 emissions were described in the previous section. A dramatic conclusion from the previous section is that opinions differ widely on where to set the target. The wide difference of opinions can arise from different judgments about the severity of climate impacts and about the cost of reducing emissions. It can also arise from different judgments about political feasibility. Other targets focus on an intermediate result which is thought to be helpful in reducing CO2 emissions. An example is the recent call by President Bush (2007) for renewable and alternative fuels to displace 15% of projected gasoline use by 2017. A similar target has been issued by the EU; it calls for 10% of cars and trucks in the EU nations to be running on biofuels made from plants. Another vehicles target has been proposed by the Governor of California. It would “apply the world’s ﬁrst low-carbon fuel standard to transportation fuels sold in California, with the goal of reducing the carbon content of passenger-vehicle fuels in the state by at least 10 percent by 2020” (CARB, 2007, p. 3). An important target in the electricity sector is the mandated fraction of electricity generation from renewable sources. In the United States, such mandates are known as Renewable Portfolio Standards (RPS). The states have taken the lead in establishing RPS goals in the United States (Ford, 2007). Renewable electricity generation is discussed in a chapter 12 in this volume. For this chapter, an important observation is that policymakers are inclined to adopt a combination of incentives to encourage compliance with the RPS. The incentives include ﬁxed payments (used by many nations in Europe), production tax credits (used in the United States), and market-based incentives such as Renewable Energy Credits (United States) or the Tradeable Green Certiﬁcates (Europe). Although ﬁxed payments are the incentive of choice in the European nations with the greatest investments (Fouquet et al., 2005), the EU’s future approach for the harmonious promotion of renewable generation is still under discussion. Regulations have also been adopted to deal with growing emissions. Some regulations deal directly with CO2 ; others deal with energy. Energy regulations can take the form of building codes, fuel efﬁciency standards on vehicles, and outright bans on inefﬁcient products (i.e., the ban on incandescent light bulbs in Australia and Canada). Regulations also deal directly with CO2 emissions. An important example is the performance standard proposed by the CPUC (2007). It would require all new long-term commitments for base load generation to serve California consumers with power plants that have emissions no greater than the emissions from a combined cycle gas turbine plant (i.e., 1100 pounds of CO2 per MWH). This rule effectively prohibits investment in new conventional coalﬁred power plants to serve base loads in California. The standard is viewed as a “nearterm bridge” until an enforceable load-based GHG emissions limit is established and in operation. A somewhat similar regulation has been proposed at the national level in the Clean Coal Act by Sen. Kerry (D-Mass.). It would ban construction of new coal-ﬁred power plants that do not have technology for carbon capture and sequestration. It is likely that many nations will adopt a combination of targets, regulations, and incentives to reduce CO2 emissions. Incentives are viewed as the crucial ingredient in order to “put a price on carbon.” That price could take the form of a carbon tax or a market price that emerges from a cap-and-trade policy. The relative merits of a carbon tax and a carbon market are discussed in the next section. It explains that both policies can deliver emissions reductions in cost-effective manner if the costs of emissions reductions

EBL

Global Climate Change and the Electric Power Industry

521

are well known. However, with uncertainty in costs, most economists prefer the carbon tax. But political factors tend to favor the adoption of carbon markets. 3

14.9. A Carbon Tax or a Carbon Market? Proponents of a carbon tax argue that it makes more sense to guarantee the cost than to guarantee that emissions will not exceed the target level. They prefer the tax because they believe it could be designed to send a predictable incentive for emissions reduction. Portney (2005) gives a simple illustration. He starts with a tax of 5$/MTC which he estimates would add around 1 cent/gallon to the cost of gasoline. The tax would then be raised by $5 every 2 years, reaching 55$/MTC at the end of a 20-year period. He argues that the tax would bring substantial advantages because of the increased revenue (estimated around $75 billion annually at the end of the 20 years). This new source of government revenue could help the nation deal with pressing problems such as the budget deﬁcit, and it could also be used to lower taxes on work or savings. The tax would also lessen the nation’s use of imported fossil fuels, thereby helping with the nation’s security problems and its trade deﬁcit. A steadily growing carbon tax would send an increasingly strong signal to motivate the investment in carbon-free technologies. Proponents argue that a predictable signal is crucial if investors are to make the commitment to carbon-free technologies needed in the future. They argue that the market approach could well lead to highly volatile prices of carbon allowances, leaving investors uncertain about future investments. Indeed, extreme volatility in emissions markets could lead to interruptions or closure. Cooper (2005, 2007) explains that most economists would be skeptical of a cap on emissions that would warrant any price for allowances to bring emissions to below the cap. He argues that expanding strict, quantitative targets to countries like China and the United States would be difﬁcult, and he believes the allocation of allowances would be a highly political, sometimes corrupting process (unless the allowances are issued at auction.) He favors a uniform tax on carbon emissions, and he argues that the tax would be economically efﬁcient, since reductions will be greatest where the cost of reducing emissions is the least. He gives an example of an initial tax of 14$/MTCO2 , but he does not speak of an upward trajectory. Rather, he suggests that the nations review the tax rate after 10 years to evaluate the impact of the tax and the change in the climatic system. Both a carbon tax and a carbon market can be designed to achieve similar emission reductions (Pizer, 1999). A cap-and-trade policy is appealing because the cap on emissions is clearly deﬁned, and the policy aims to guarantee that the emissions do not exceed the cap. If the costs of emission reductions are known, a carbon tax could be designed to achieve the same reduction. (The carbon tax would turn out to be the same as the price that emerges from the carbon market.) Pizer explains that the two policies differ markedly when we consider the uncertainty in the cost of emissions reduction. His analysis considers a carbon tax of 80$/MTC, which he estimates to be equivalent to a cap of 8.5 GTC of worldwide CO2 emissions by the year 2010. He introduces uncertainty in the

EBL

3 This chapter discusses the choice between a carbon tax and a carbon market. But the discussion does not have to be framed as an either-or question. A nation might adopt a carbon tax as well as cap-and-trade legislation. This idea was suggested by former Vice President Al Gore in Testimony before the US Senate on 21 March 2007.

522

Competitive Electricity Markets

uncontrolled emissions and the costs to reduce emissions. He then conducts simulations to learn the range of possible outcomes with either the tax or the market. The simulations revealed that the carbon tax produced net gains ﬁve times higher than even the most favorably designed carbon market. They also revealed a narrower range of impacts: from 0 to 2.2% of gross domestic product (GDP) for the market, from 0.2 to 0.6% of GDP for the tax. Pizer acknowledges that a carbon tax faces steep political opposition in the United States. Businesses tend to oppose the transfer of tax revenues to the government, and they may hold out hope that the many of the allowances in a carbon market would be allocated for free. Some environmentalists oppose the tax because it might not be varied over time in such a manner as to guarantee a reduction in emissions. A political path forward could involve a carbon market subject to a penalty price. The penalty price is a preset price, and the market operator offers to sell unlimited numbers of permits at that price. It is sometimes called the threshold price, safety valve price, or the price cap. The price cap places an upper limit on the market clearing price, so there is some protection against soaring prices. Setting the price cap is difﬁcult. If the price cap is set extremely high, participants are afforded little protection against soaring prices, and the market is vulnerable to failure. If the price cap is set too low, participants will often be paying the penalty price rather than making the costly investments needed to lower emissions. The problem in setting the price cap is that there are widely differing estimates on carbon allowance prices that will be needed to comply with the emissions goals (Fisher and Morgenstern, 2006). Another problem is that policymakers may set a low price cap to help gain legislative approval. An example of this approach is the National Commission on Energy Policy proposal for cap-and-trade with a low safety valve price. Their cap would be set at 7$/MTCO2 in 2010 and grow at 5%/yr in nominal dollars. This was recommended because the commissioners believed “that reducing uncertainty and likely opposition by explicitly capping program costs and impacts is the best path toward timely action” (NECP, 2004, p. 23). A few years later, the NCEP (2007, p. 5) called for a somewhat higher price cap. It would start at 10$/MTCO2 and grow at 5%/yr in real dollars. The economic arguments for carbon taxes over carbon permits are reviewed by Jacoby and Ellerman (2004). They conclude that a carbon tax is preferable when dealing with highly uncertain costs and a stock pollutant like CO2 . [In testimony before the US Senate, Jacoby (2007) explained that “among the market-based approaches, a universal national carbon tax is the favorite of many economists.”] They then discuss the idea of a carbon market designed with a low price cap (a cap that would frequently set the market price of carbon). In this situation, emissions would remain above the target, and there could be a lack of commitment to raising the price cap sufﬁciently to allow the emissions goal to be achieved. The inability of the market to meet emissions targets would be a continuing source of controversy, bringing into question the use of the carbon market to control emissions. They summarize this approach as “little more than a band-aid on an inappropriate implementation of cap and trade.” Although their reasoning favors the carbon tax, they are aware that “the dominant choice seems still to be a carbon market, not only in the US but also in Europe and elsewhere.” A different perspective on the advantage of a carbon tax is provided in computer simulations by Fiddaman (2002). His model simulates economic activity, emissions, carbon prices, atmospheric CO2 , and temperature impacts. The most striking result is the extreme volatility in carbon prices with Kyoto targets on emissions. (The simulations show the possibility of carbon prices soaring to almost 1000$/MTC shortly after the market opening.)

EBL

Global Climate Change and the Electric Power Industry

523

The volatility of a carbon market is a serious problem, both for investors and the political viability of the permit system. Fiddaman believes that carbon taxes are the preferred policy, but he is aware of the unpopularity of taxes and the need for prompt action to reduce anthropogenic emissions. His article closes with a highly pragmatic recommendation: The challenge for policymakers is to take aggressive action without causing extreme short-run economic disruption, triggering a political backlash. For this purpose, taxes appear to be a more suitable mechanism than emissions permits. However, since further delay is in itself expensive, it seem sensible to make modest changes to current agreements and take ﬁrst steps as soon as possible; permits can always be scrapped for taxes later. (Fiddaman, 2002) Carbon taxes have been implemented in some countries (i.e., Sweden, Finland, and Norway), and they have received vocal support from some prominent individuals in the United States. But political considerations favor the adoption of carbon markets. The term tax is a “dirty word” in many political discussions in the United States, and the adoption of a tax in the EU would require a unanimous vote of the member countries. (A majority vote is sufﬁcient to adopt a carbon market.) The EU has moved ahead with a carbon market, and the proposals in the United States all involve different versions of a carbon market. The preference of US policymakers for carbon markets is inﬂuenced by the nation’s previous experiences with energy taxes. Taxes have not been popular in the United States, and the country’s experience with energy tax credits has also proven to be unreliable. The “on-again, off-again” experience with the renewable energy production tax credit is an example (Sterzinger, 2006). These experiences leave the impression that it will be difﬁcult to implement a carbon tax and ensure that the tax will remain in effect over time. And ﬁnally, some ﬁnd it difﬁcult to envision how investors could count on a carbon tax to follow a preset upward trajectory. The preference for carbon markets in the United States is also inﬂuenced by the prior experience with market-based control of emissions, which is summarized in Box 14.8.

EBL

Box 14.8. United States’ experience withemissions trading Market-based control of pollutants in the United States includes control of lead, nitrogen oxides, and sulfur dioxide. The sulfur dioxide program is often cited as an exemplary example of the beneﬁts of trade in allowances for air pollution because it is said to have achieved important reductions at costs far lower than what might have been achieved under previous programs (sometimes referred to as “command and control” programs). This perceived success is due largely to the unexpectedly low costs of sulfur allowances. But the low costs could be due to fortunate market conditions as well as a good market design, as noted by Ellerman et al. (2000, 2003) and Reilly and Paltsev (2005). Since these reports were published, the SO2 market has shown surprisingly volatile prices in the wake of the Clean Air Interstate Rule. Analysts expected the market to respond to the new rule with current vintage SO2 allowances selling for around $600 per ton. The price was under $200 in 2003, and it climbed to $600 by the end of 2004. It then soared past the expected value and peaked at $1600 by the end of 2005. It fell precipitously in early 2006 reaching

524

Competitive Electricity Markets

$600 midway through 2006 (EPA, 2006, p. 9). Clearly, this market warrants further observation. The key lesson from the emissions trading in the United States is that trading works best when the allowances are clearly deﬁned and tradable without case-bycase pre-certiﬁcation (Ellerman et al., 2003). Early advocates of emissions trading in the United States emphasized the gains from trading among participants. For example, companies with low costs to reduce emissions would achieve greater reductions and sell allowances to companies facing much greater costs. The experience from the acid rain program indicates that intertemporal trading can also help to lower the total costs of compliance. Advocates normally encourage “banking” of allowances, in which companies reduce emissions ahead of the target date and “bank” the allowances for use in future periods when compliance is more expensive. Additional information on the US experiences is given in Markets for Clean Air (Ellerman et al., 2000), in the Pew Center report on Emissions Trading in the US (Ellerman et al., 2003), and in the appendix to the CARB (2007) report of the market advisory committee. Advocates of carbon markets are aware of the concern that market prices could be highly volatile; especially if caps on emissions are lowered quickly before new investment in carbon-free technologies have time to respond. The concern over volatility is a legitimate concern, given that our experiences with emissions trading are mixed (Ellerman et al., 2003). However, there are differences of opinion on whether banking would be allowed (Reilly and Paltsev, 2005, p. 16). Absent banking, the penalty price is the main defense against extreme price volatility

EBL

Although many policymakers are calling for cap-and-trade in emissions, the United States has not yet committed to a mandatory market in carbon emissions. A survey of power industry executives (CERA, 2006) found that 80% expect binding carbon caps to start within the next decade. However, the nation’s current experience is limited to the voluntary market operated by the Chicago Climate Exchange (CCX), a voluntary, but legally binding greenhouse trading system. It provides a testing ground for participating companies to reduce emissions and help shape the rules and procedures for emissions trading. The CCX is helpful, but the ﬁrst serious test of a carbon market is the Emissions Trading Scheme (ETS) in Europe. 14.10. The European Emissions Trading Scheme The ETS was established in 2003 by directive of the European Commission (EC, 2003). It is to run for a 3-year mandatory “warm-up” period in which CO2 permits are traded among participants from six industry sectors. The goal is to help prepare EU member states to achieve compliance with their international commitments from 2008 to 2012 under the Kyoto protocol. The next step is a 5-year phase which could include additional sectors of the economy and the trading of other GHGs. This is the world’s ﬁrst, serious test of mandatory cap-and-trade in carbon, so it is widely watched. Current news of the ETS is provided by Point Carbon (www.pointcarbon.com). Reports on the ETS experiences are available from the Pew Center (2005), the World Bank (2006), the CARB (2007), and from MIT (Reilly and Paltsev, 2005). This summary draws heavily from the MIT analysis.

Global Climate Change and the Electric Power Industry

525

The ETS was designed as a test system with a relatively mild requirement for emissions reduction at the start. The idea was to give participants the opportunity to become familiar with the basic operations of establishing registries and the trading instrument. At the same time, the member countries were negotiating National Allocation Plans (NAPs) with the EC. The overall effect of the NAPs is difﬁcult to estimate, but the general goal was to limit growth in emissions, which could correspond to an emissions reduction around 1% below a reference case (Reilly and Paltsev, 2005, p. 2). An important decision at the outset was whether participants receive allowances for free or whether they should purchase them at auction. Such a decision was the subject of much debate and negotiation in the creation of the US Acid Rain Program (Ellerman et al., 2000). Some argued that fairness required allowances to be sold at auction rather than given out for free. Others argued that an auction was not politically possible. If this is true, the pragmatic question is how to allocate free allowances among the many participants. Ellerman et al. (2000, p. 36) estimated that this question was equivalent to the “handing out between $45 and $63 billion” in the US Acid Rain Program. He commented on the politics of the allocation: “with that sort of rent on the table, one would certainly expect to see serious rent seeking, and Washington did not disappoint.” In the end, the SO2 allocation was determined approximately by a generator’s historical fuel consumption (millions of BTUs/year) multiplied by a standard emissions rate (i.e., 2.5 pounds of SO2 /million BTU), with the actual allowances scaled to comply with the cap. Allocating emissions based on the input fuel to power plants is one of several allocation approaches, some of which are described in Box 14.9.

EBL Box 14.9. Alternative methods to allocate free allowances Allowances may be sold at auction or given away free. Most economists “would favor the auction approach when allocating among nations” according to Cooper (2007). But he observes that “auctions are in fact rarely used when valuable resources are to be distributed.” Most participants in a recent workshop on carbon markets (NECP, 2006) viewed auctions as “a politically impractical approach for distributing a majority of allowances in the near term” (NECP, 2006). So, it appears likely that most of the CO2 allowances will be allocated at no cost to the participants at the opening of carbon markets. The allocation is likely to be a highly political process, with each step in the process subject to debate. The political debate may well begin with the very ﬁrst step – the selection of a reference year (i.e., the year 1990). A year with unusually high emissions would be favored by those seeking to minimize the need to reduce emissions, as discussed by Cooper (2006, 2007). The emissions in a reference year (or averaged over a reference period) are then compared to the target reduction to establish a cap on total emissions. In multinational programs (such as the EU), allowances are allocated among nations. Within each nation, allowances must be allocated among industries. Within each industry, allowances are allocated among installations. Allocating allowances among nations could depend on much more than the nation’s CO2 emissions during the reference period. It might also depend on whether the nation is an early or late entrant into the market, and it could also depend on the

526

Competitive Electricity Markets

extent of economic development and the size of the population. For example, the “convergence” proposal calls for carbon allowances to be allocated so that per capita emission rights are equal across nations (Meyer, 2000). Allocating allowances to different industrial sectors could also depend on more than each sector’s emissions during the reference period. For example, some industrial sectors may be viewed as crucial to the nation’s exports, and they would be favored with larger share of the allowances. Also, some sectors might be in a better position to capture windfall proﬁts from free allowances, so their share of free allowances might be reduced. Allocating allowances among installations could be based on historical use of carbon fuels. In an “input-based” system like the US Acid Rain Program, for example, the emissions might be based on the amount of fuel input used by power plants during the reference period. An “output-based” allocation of allowances would use the allocation process to beneﬁt investors who have committed to technologies with low carbon emissions. The approach is explained in a short paper by Harmon and Hirschhorn (2006) and a detailed report by Bird et al. (2007). An output-based allocation would grant allowances to electricity generators based on their fraction of total electricity demand, not on their fraction of total CO2 emissions. This allocation would apply to all generators regardless of the source of fuel, so it would provide a boost to carbon-free generators. For example, wind companies would beneﬁt since they would receive allowances but have no obligation to turn them in. They could then sell the allowances to coal-ﬁred generating companies which face an obligation to surrender allowances at the end of the accounting period. Some wind companies would value the output-based allocation as a way to ensure that a voluntary market for “green power” would not be eliminated by a mandatory carbon market (Bird et al., 2007; Harmon and Hirschhorn, 2006). The wind companies serving voluntary customers could choose to remove their allowances from the market. This would reduce the total CO2 emissions and provide the basis for their claim that they are able to “deliver green power” to their voluntary customers. Whatever the allocation method, the receipt of free allowances can raise potentially troublesome questions for regulated utilities. Participants in a NECP (2006) workshop anticipated that many regulated entities could end up with enough allowances to cover most of their emissions. At the same time, they might be able to pass the opportunity costs of all allowances used for compliance through to consumers. The regulated entities would then be more proﬁtable after the implementation of a mandatory carbon market. The case of “windfall proﬁts” could present political problems for the continuation of the market. This problem is identiﬁed by the CARB market advisory committee which estimates that the electricity industry in the United Kingdom enjoyed windfall proﬁts of E500 million in the ﬁrst year of the ETS alone (CARB, 2007, p. 52). News stories of windfall proﬁts in Europe (Fairﬁeld, 2007; Samuelsohn, 2007) have led ofﬁcials from the United States to question the ETS decision to allocate almost all of the allowances for free. According to a New York ofﬁcial involved with RGGI, “the way we distribute the allowances will be vastly different than the European experience” (Fairﬁeld, 2007). The next phase of the ETS may be quite different as well. According to the CARB (2007, p. 97), Spain and other member states are reducing the free allocation to the electricity sector to reduce the windfall proﬁts to this sector.

EBL

Global Climate Change and the Electric Power Industry

527

The ETS was ofﬁcially launched in January 2005. The allowances were allocated among nations based on historical emissions with some adjustment to account for the change in emissions due to economic restructuring. (The purpose of this adjustment is to avoid granting allowances well above emissions, thus creating what is sometimes called “hot air.”) The allowances were then allocated among six industry sectors, with 95% of the allowances issued for free. It was initially estimated that around 5000 installations would be affected, but more than 12 000 installations were affected by the time the commission ﬁnished interpreting which facilities were obliged to participate. A main purpose of the ETS is to “put a price on carbon.” So the price behavior during 2005–6 drew considerable attention in the MIT report. The market opened with prices under 10E/MTCO2 . (The remainder of this section drops the MTCO2 for brevity.) Prices grew to E29 by July and were trading around E20–22 during the fall of 2005. Spot prices rose again early in 2006 and exceeded E30 by April 2006. These prices were unexpectedly high according to several estimates. For example, the International Energy Outlook of the EIA (2006, p. 76) estimated an upper bound of $42 for a scenario with Kyoto protocol goals remaining in place over the next two decades. At 1.27$/E, the EIA estimate corresponds to E33. The surprising result is that the ETS price rose nearly to this level during a “warm-up” period in which the participants were supposed to be operating with a relatively small requirement. Understanding the reason for the surprise is made difﬁcult by a variety of confounding factors. They include allocation of allowances, changes in the weather and fuel prices, expectations about the contribution of CDM, and uncertain economic growth, especially the growth in the eastern European countries. The MIT report compared the ETS prices with expectations from several studies. For example, a Pew Center study placed the likely carbon price at E5.50 with a range from E2.50 to E10. The early ETS prices were ﬁve times higher than the Pew Center’s median estimate. The MIT report also referred to the modeling forum, an important way to compare projections across a range of different models. The forum study (Weyant and Hill, 1999) reported that seven of eleven models showed carbon prices around E25. This price corresponds to the prices seen in the early months of the ETS, but the forum participants were working with a scenario calling for 20% reduction in emissions below a reference case (a much more ambitious goal than the 1% reduction in the NAPs). So the puzzle of the high prices remained. The MIT report then turned to the MIT’s Emissions Prediction and Policy Analysis (EPPA) model to estimate carbon prices under conditions similar to the ETS. The results led to an even more puzzling situation: the market clearing price turned out to be less than E1. The prices in the early months of the ETS are unusually high when compared to projections from many of the multi-sector general equilibrium models. But these prices would not be surprising to participants from the electricity sector. The electric power sector is one of more important of the six sectors participating in the ETS. Europe has a substantial amount of coal-ﬁred generation, and the coal plants face variable costs higher than nuclear but lower than gas-ﬁred combined cycle plants. As the price of carbon allowances rise, the coal plants’ variable costs increase dramatically because coal is the most carbon intensive of the fossil fuels. The gas-fueled combined cycle (CC) plants also face increasing variable costs, but the increase is less severe because of the higher efﬁciency of combined cycle units and the lower carbon content of natural gas. When carbon prices rise sufﬁciently high, generating companies are inclined to cut back on the operation of the coal plants and compensate with greater use of the gas-fueled CC plants. The point at which this fuel switching occurs depends on the price of natural gas

EBL

528

Competitive Electricity Markets

and the relative efﬁciencies of the gas and coal units. With gas prices between $5 and $6 per million BTUs, for example, a carbon price around $36 might trigger some fuel switching. The corresponding fuel-switching price in the ETS would be around E28. Thus, one possible explanation of the early ETS prices is that participants were expecting fuel switching in the electricity sector to be the dominant method of achieving the emissions reductions. A study by de Leyva and Lekander (2003) supports the fuel-switching explanation. They used a proprietary McKinsey model of the European electric power industry to arrive at carbon prices around E15 in 2005 and E25 by 2008. Fuel switching in the electricity sector was also the major assumption in the 2006 Climate Change Program in the United Kingdom. They projected that allowance prices would average around E20 during 2006–7, and that the impact of the ETS “is expected to be felt most strongly in the power generation sector where there is considerable scope for CO2 abatement through switching fuel inputs” (UK Energy and Environment, 2006). The McKinsey and UK projections provide a way to make sense of the initial ETS prices, provided one believes that the principal response of the ETS participants is fuel switching in the electricity sector. However, the ETS situation changed dramatically in May 2006 when spot prices fell abruptly to less than E10. This threefold decline coincided with the release of a registry report that veriﬁed the participants’ emissions. The release of the registry report revealed that the allocated allowances exceeded the veriﬁed emissions. In the months following the abrupt decline, spot prices climbed to around E15 and remained at that level through the summer of 2006. The ETS price movements during 2005 and 2006 are reported here in detail because of the surprising volatility during a “warm-up” period in which low prices were expected. The initial prices were viewed as far too high by many experts. After the price decline in May 2006, the prices were viewed as far too low to support investment in emissions reduction. Whatever their viewpoint, all observers would agree that the early prices conﬁrm the warnings that carbon markets could expose participants to highly volatile prices. Reilly and Paltsev (2005) view the creation of the ETS as a “watershed event in climate policy.” They believe the performance of the ETS may well determine whether there is rapid progress toward establishing international markets in permits or whether the world turns sour on permit trading and pursues other policies. They argue that economic theory strongly concludes that cap-and-trade systems for controlling pollutants can achieve emissions reductions at lower costs than traditional programs (of the “command and control” variety). But if cap-and-trade is to be adopted more widely, policymakers need to know how to learn from the ETS experience. One lesson is that the encouraging results from the US Acid Rain Program may lead to unrealistic expectations for the ETS. They describe the “near-legendary status” of the acid rain trading program among some in the environmental community because “it was perceived to reduce the cost of abatement by an order of magnitude.” The claim of an order of magnitude reduction is based partly on “some exaggerated early cost projections and some fortunate circumstances unrelated to emissions trading.” The ETS experience is “exactly the opposite of that with sulfur trading in the United States. The permit trading price is an order of magnitude higher than what was expected.” Some might draw the conclusion that the ETS has failed, but they view this as an unfortunate and probably unwarranted conclusion. Certainly, the ETS prices were a “surprise (if not a shock) based on expectations that reductions required would be

EBL

Global Climate Change and the Electric Power Industry

529

very mild.” The MIT report concluded with a call for patience4 in drawing conclusions on the ETS. The remainder of this chapter focuses on the electric power sector. The power sector is at the center of many state and regional proposals in the United States and it is one of the key sectors in the current test of the ETS. The response of the power sector is best explained with a concrete example of carbon market legislation. 14.11. S139 and the Response of the US Electric Power Industry S139 was introduced by Senators McCain and Lieberman in January 2003. It did not pass, but it was the subject of detailed studies by MIT (2003) and by the US Energy Information Administration (EIA, 2003). These studies are helpful in showing the key role of the electric power sector. This chapter draws heavily on the EIA analysis. Starting with the macro-economic impacts, the EIA found that the nation could reduce GHG emissions to meet the S139 targets with a small impact on the economy as a whole. The EIA expected GDP to grow 3.04%/yr over the next 20 years in a base case scenario. With S139, the nation’s GDP was projected to grow at 3.02%/yr (EIA, 2003, p. 206). When this reduction is accumulated over a 50-year period, the total impact is somewhat less than a 1% reduction in GDP. This estimated impact for the United States matches estimates of worldwide GDP impact from the Stern Review. It is also generally consistent with the estimated impact in the IPCC’s Third Assessment. Working Group III on Mitigation found that impacts by the year 2010 would range from 0.1 to 1.1% of GDP in scenarios with full emissions trading between Annex B countries. A 1% reduction in GDP is an important impact. In the United States, for example, 1% corresponds to around $120 billion per year or around $400 per year per person (Lohr, 2006). The power industry is expected to deliver a major share of the emissions reductions because of the many technologies for generation and because electricity accounts for an important share of CO2 emissions. Electricity is responsible for 39% of energy-related emissions and 33% of total emissions in the United States. Coal-ﬁred generation provided about half of the nation’s generation in the year 2000 (Ford, 2006). Figure 14.8 summarizes the key results from the EIA analysis of the nation’s electricity sector. Emissions are shown on the left scale. In a business-as-usual case, electricity sector emissions were predicted to reach 868 MMTC by the year 2025, 40% higher than in the year 2000. The cost of carbon allowances are shown on the right scale. Finding these prices was a complicated task involving a large collection of models. The EIA used repeated simulations of the many models to search for an allowance price trajectory that would induce the entire economy to achieve the emissions goals. Their search led to a price

EBL

4

The call for patience is important, but it may be difﬁcult to heed. Recent news stories (Samuelsohn, 2007; Fairﬁeld, 2007) reveal an urgent desire by US ofﬁcials to craft carbon market proposals that are informed by the ETS experience so far. In California, the market advisory committee for the CARB (2007, p. 97) views the ETS to be a success (in light of its tight implementation schedule). The committee summarizes lessons from the ETS and recommends that CARB create a cap-andtrade system that would avoid the ETS price volatility. The key measures to reduce volatility are to allow banking of allowances and to delay the initiation of market operation until monitoring and enforcement measures are ﬁrmly established. (The safety valve, a common measure to limit price volatility, is to be avoided, according to advisory committee. The carbon tax, an entirely different approach to put a more predictable “price on carbon” is dismissed (in three sentences) as not suitable given the AB32 goal for emissions reduction.)

Competitive Electricity Markets

530

90

80

2

86

8

900

80

759

Price to achieve compliance in the entire economy

69

9

Base case emissions

62

1

700

70

Phase I = 621 MMTC

50

48

5

61

5

Phase II = 492 MMTC

60

500

60 S. 139 emissions

50

MMTC

600

400

40

35

300

30

200

20

20

5

22

100 0 2000

$ / MTCO2

800

10 0 2005

2010

2015

2020

2025

Fig. 14.8. Summary of expected impacts of S139 on the US electricity sector. Source: Adapted from information in the Energy Information Administration (EIA, 2003) report on S139.

EBL

somewhat above 20$/MTCO2 when the market opens in 2010. By the year 2025, the price of carbon allowances would reach 60$/MTCO2 .5 The EIA estimated that emissions would be reduced dramatically. Indeed, the electricity sector emissions would decline well below its allocated allowances. The allocated allowances are 621 MMTC in phase I and 492 MMTC/yr in phase II. There is much debate over whether allowances should be allocated for free or purchased at auction. The free allocation is often justiﬁed as a “grandfathering” provision on the grounds that fossil-fueled facilities were constructed in an era when the climate change problem was not evident. The S139 legislation called for 80% of the allowances to be allocated for free at the start of the market; 20% would be purchased at an auction run by the non-proﬁt corporation administering the market. The mix would change over time as industries are expected to become familiar with the market and with their obligation to reduce carbon emissions. By 2025, 20% of the allowances would be free and the 80% would be purchased at auction. Figure 14.8 shows that the electricity sector emissions would be well below the allocated allowances. In 2015, for example, the electricity sector would have around 120 MMTC of extra allowances that could be banked for future use or sold to less responsive sectors in the economy. With the transition to phase II at hand, there is a strong need for banked allowances. By 2025, the industry would have 287 MMTC of allowances that would probably be sold to other sectors which face a more difﬁcult job in cutting emissions. The upward price trajectory in Fig. 14.8 is the result of iterative calculations of many different models to ﬁnd the price that would induce the simulated participants to bring emissions into compliance with S139. The smooth price trajectory may leave the impression 5 These prices are quite close to the previously mentioned MIT (2007b) study of the “50% reduction” proposals before the US Congress.

Global Climate Change and the Electric Power Industry

531

that there is little potential volatility in the market proposed by S139. This would be an improper interpretation. The EIA calculations do not deal with some of the market situations discussed previously in the ETS. Also, the calculations assume extensive use of banking. The banking is quite important at the transition year 2016 (where we see no discernible change in the upward price trajectory). In the absence of banking, however, the price trajectory would be quite different. The EIA estimates that the phase I carbon prices would be considerably lower. They would then jump ﬁvefold during the transition to phase II. Carbon prices were projected to reach 75$/MTCO2 during the ﬁrst year of phase II. Prices were then expected to decline over the next few years before ﬂattening out at a value slightly below the ﬁnal price of 60$/MTCO2 shown in Fig. 14.8. In both cases, the price trajectories are sufﬁciently high for the nation to achieve the emissions goals speciﬁed in S139. The comparison of these price trajectories adds support to those who believe banking of allowances is important to help moderate the volatility in carbon prices. Turning to emissions reductions, the EIA found that the electricity sector would achieve a 76% reduction in carbon emissions by the year 2025. This could be achieved with a 46% increase in the average retail electricity rate charged in the year 2025. Figure 14.9 puts these main results in perspective by showing CO2 reduction on the vertical axis and price increases on the horizontal axis. The graph is divided into diagonal halves by a 50/50 line to show which sectors would be most responsive under S139. The idea behind cap-andtrade is that market forces will bring forth a strong response from those sectors with the greatest ﬂexibility in response. Less ﬂexible sectors would then buy the needed allowances from the more responsive sectors. The transportation, industrial, and residential sectors are simulated to be the less responsive sectors. Their impacts fall well below the 50/50 line. The electricity sector is well above the line. It is expected to lead the way in reducing carbon emissions. Figure 14.10 shows the simulated response within the electricity sector that makes the 76% reduction in emissions possible. The horizontal axis is measured in billions of kwh/yr of change in generation. The four black bars at the bottom of the bar chart show the

EBL

100

Percent reduction in CO2 emissions

80

The USA electricity sector (EIA)

60 The 50/50 trend line serves to separate the results. The unresponsive sectors fall below the line.

40

The electricity sector is the most responsive. It should be above the line,

20 Industrial (EIA)

Transportation (EIA) Residential (EIA)

0 0

20

40

60

80

100

Percent increase in the retail price of energy

Fig. 14.9. Summary of expected impacts of S139, the Climate Stewardship Act of 2003. Source: Adapted from information in the Energy Information Administration (EIA, 2003) report on S139.

Competitive Electricity Markets

532

New coal w/Seq. New nuclear New gas CC w/Seq.

More geothermal More wind More biomass Demand reduction

Less gas CCs Less CTs Less fossil steam Less regular coal 0

200

400

600

800

1,000 1,200 1,400 1,600 1,800 2,000 2,200 2,400

Change in generation (billions of kwh) in the year 2025 Fig. 14.10. Summary of the expected responses to S139 in the US electricity sector. Source: Author’s adaptation of information in the EIA (2003) report on S139.

reduction in fossil-fueled generation caused by the allowance prices in S139. The dominant response is clearly the large reduction in conventional coal-ﬁred generation. The remaining bars in Fig. 14.10 show the estimated changes to compensate for the reductions in the conventional, fossil-fueled power stations. The ﬁrst mode of compensation is a demand response to the higher retail prices. The EIA estimates that demand would be reduced by nearly 600 billion kwh/yr by the year 2025, an 11% decline compared to the reference case. The increased response from renewable generators (biomass, wind, and geothermal) are shown next. (Renewable generation is described in a separate chapter in this volume.) The EIA expected generation from biomass-fueled plants to be the most responsive of the renewable technologies, providing an extra 400 billion kwh/yr by 2025. Wind generation is projected to provide only 0.6% of total generation in the reference case. The EIA projects that wind could account for 6% of total generation in the S139 case. There is also a small contribution from extra geothermal generation. The top three bars in Fig. 14.10 draw attention to the advanced technologies that require further research, development, and/or policy changes to ensure their response. The most responsive of these technologies are expected to be new gas-fueled CCs with carbon capture and sequestration equipment. The EIA cautions that any of the advanced technologies could emerge in a S139 scenario depending on relative prices of coal and gas and the progress in research and development. With the EIA assumptions, gas-fueled CCs are the dominant use of sequestration technologies. However, new coal plants with carbon capture and sequestration are also projected to provide additional generation in the S139 case. The EIA projects that there would also be increased investment in new nuclear generating stations in the S139 case which would provide over 400 billion kwh/yr of additional generation. Figures 14.9 and 14.10 provide a visual summary of expectations that many hold for the near-term response of the electric power industry to carbon market legislation. The diagrams show that the power industry is likely to achieve far greater emission reductions

EBL

Global Climate Change and the Electric Power Industry

533

than other sectors of the economy. The industry will achieve this strong response largely through reduction in conventional coal-ﬁred generation. The electric sector is expected to compensate for the reduced coal generation by a price-induced demand reduction, increased generation of existing renewable technologies, and increased generation from advanced technologies. The main conclusion from Fig. 14.10 is that contributions from many technologies will allow the electricity sector to lead the way in reducing CO2 emissions. An important question for current policymakers is the wisdom of initiating carbon markets if the electricity sector response depends on the increased generation from the advanced technologies shown at the top of Fig. 14.10. Carbon sequestration technology has great potential (IPCC, 2005; MIT, 2007a), but it is not commercially available. Advanced nuclear technologies could provide an alternative source of carbon-free generation, but the industry currently faces stagnation and decline. If new nuclear plants are to provide signiﬁcant additional generation, the industry must deal with the unresolved problems of high construction costs, waste disposal, proliferation, security, and safety. The challenges of the nuclear industry are not described here. [The reader is referred to the MIT (2003) report on The Future of Nuclear Power.] The challenges of carbon capture and sequestration are also left for others to describe (IPCC, 2005; MIT, 2007a). The fundamental question to be considered here is whether the United States should adopt carbon legislation or postpone action until the advanced carbon-free generating technologies become available. Perspective on this question is provided by a recent study of the impact of S139 on the electric power system in the western United States and Canada.

EBL

14.12. S139 and the Near-Term Response in the Western United States and Canada Electricity generation in the western United States and Canada is provided in a large, interconnected power system known as the WECC, the Western Electricity Co-ordination Council. The boundaries of the WECC and its four power areas are shown in Fig. 14.11. This region has considerably more hydro resources, and it makes less use of coal-ﬁred generation than the nation as a whole. The impact of S139 on the WECC was the subject of a recent study to learn if the western system could achieve the large reductions in CO2 emissions that were projected by the EIA

I

II

IV III

Fig. 14.11. Map of western states and provinces in the WECC.

Competitive Electricity Markets

534

for the nation as a whole. The study adopted conservative assumptions on the advanced technologies of carbon sequestration and new nuclear plants. For example, gas-fueled CCs with carbon sequestration were selected as the most likely form of carbon sequestration. With gas prices at $5.50 per million BTUs, the investors in CCs with sequestration were assumed to face a total levelized cost of around 75$/MWH (Ford, 2006). This technology was the only advanced, carbon-free technology, so it served as the “backstop” technology in the modeling calculations. But it would not become available until 2020, and the CCs with sequestration would provide only a small response to the carbon prices from S139. This means the key to the simulated WECC response lies in a combination of demand reduction programs and increased response from existing renewable technologies such as biomass and wind. Many states are encouraging newer, more aggressive efﬁciency programs as part of the effort to lower CO2 emissions. The WECC study adopted the conservative assumption that only the current efﬁciency programs would remain in effect. With these assumptions, biomass and wind generation are the keys to the near-term response in the western system. The west has huge wind resources (Ford, 2006), so it is important to put the EIA projections of wind generation in perspective. Figure 14.12 compares the percent of generation from wind in the EIA study with wind generation percentages and targets in Europe. (Wind and other renewable generation are described in great detail in a separate chapter.) Figure 14.12 shows the EIA results at the bottom of the bar chart. Wind is projected to provide 0.6% of US generation by the year 2025 in the reference case. With the high prices of carbon allowances, wind investors will have a strong advantage over their competitors because of the price of carbon allowances. The EIA estimates that this advantage would allow wind to provide 6% of the nation’s electricity generation by the year 2025. This is a 10-fold increase, a seemingly dramatic response to S139. But the response is less dramatic when it is compared with the wind generation in Europe.

EBL

Denmark’s goal for 2015

35.0% 20.0%

Denmark in 2004 German study for 2015

14.0%

Spain’s goal for 2011

14.0%

Spain in 2004

6.5%

Germany in 2003

4.3%

USA in 2025 with S139

6.0%

USA in 2025 (EIA)

0.6%

USA in 2000

0.2%

0%

5%

10%

15%

20%

25%

30%

35%

40%

Fig. 14.12. Comparison of various projections of wind generation. Sources: The US estimates are from the EIA (2003) report on S139. The estimates for wind generation in Europe are taken from several sources, as explained by Ford (2006).

Global Climate Change and the Electric Power Industry

535

The top bars in Fig. 14.12 put the EIA projections in perspective by showing current and planned wind generation in Europe. These and other results in Fig. 14.12 reveal that the potential for wind generation is much higher than suggested by the projections in the EIA study of S139. The cost of expanding the west’s wind capacity was simulated in a conservative manner in the WECC study. The study assumed there would be no reduction in wind costs as more capacity is constructed and operated over time (i.e., cost reductions from learning and from R & D were ignored). To be conservative, the study assumed that wind’s costs would increase as more and more capacity was added to the system. Higher construction costs were incorporated as the simulated investors turned to wind farms with longer transmission connections or in lower-class wind regions. Higher costs were also attributed to challenge of managing the growing amount of intermittent generation. A conservative approach was also adopted for biomass-fueled generation. For biomass to be considered carbon-neutral, fuel should be part of a cycle of biomass growth, harvesting, and consumption in the power plants. For the WECC study, biomass generation was based on fuel from dedicated crops (i.e., such as hybrid poplars.) To be conservative, cost reductions from the growing and transportation of the fuel were ignored. Rather, we assumed that investors would face higher and higher costs for delivered biomass as tree farms were located in less advantageous parts of the region. The purpose of these cautious assumptions was to learn the WECC’s likely response to S139 prices with existing generating technologies. Figure 14.13 portrays the simulated

EBL

250

CTs Gas steam

200 Gas fueled combined cycle

150

Shaped hydro

100 Biomass Coal

50 Nuclear Other

Wind Base hydro

0 24

1

Hours in a summer day Fig. 14.13. Projected generation for a peak summer day in 2024 in the reference case.

Competitive Electricity Markets

536 250

200

gas fueled combined cycle

150

100

shaped hydro biomass CC w storage

nuclear

50

wind other

base hydro

0

1

24

Hours in a summer day Fig. 14.14. Projected generation for a peak summer day in 2024 in the S139 case.

EBL

response by showing a typical day in the summer of the ﬁnal year of the reference simulation. Figure 14.14 shows the corresponding results in the ﬁnal year of the S139 simulation. The simulations assumed adequate transmission capacity between the different areas of the WECC, so there is no congestion in the system. The daily dispatch patterns are shown for the entire WECC. The side-by-side comparison makes it easy to visualize the change in the WECC system 15 years after the carbon market opens. A comparison of the peak loads shows that the demand for electricity would be reduced. The reduction in the WECC simulation was 9%, which is due entirely to the consumers’ reaction to higher retail electric prices over time. (To the extent that utilities response included more aggressive efﬁciency programs, the S139 case would show larger reductions in the demand for electricity.) The most dramatic difference in the comparison is the complete elimination of coalﬁred generation. Coal-ﬁred units are shown to operate in a base load mode in Fig. 14.13. They provide around 28% of the annual generation but account for around two-thirds of the CO2 emissions in the western system. The carbon prices from S139 make investment in new coal-ﬁred capacity unproﬁtable at the very start of the simulated market in the year 2010. As the carbon prices follow the upward trajectory in Fig. 14.8, it becomes economical for utilities to cut back on coal-ﬁred generation and compensate with increased generation from gas-ﬁred CC capacity.6 With the gas prices used in the WECC study, the fuel switching would push the coal units into the difﬁcult position of operating fewer and 6

This is the fuel-switching response mentioned previously in the discussion of the ETS prices.

Global Climate Change and the Electric Power Industry

537

fewer hours in a day. Eventually this short duration operation is no longer feasible, and coal generation is eliminated completely.7 Increased generation from wind and biomass capacity goes a long way to compensating for the loss of coal generation. Wind generation responds more strongly to the S139 prices, even though investors are assumed to face higher and higher costs to bring more difﬁcult wind farms into the operation (Ford, 2006) By the end of the simulation, wind is providing around 25% of total generation. This is a substantial contribution, far beyond the EIA projections. But 25% generation from wind is well aligned with the ambitious goals adopted in the European countries noted in Fig. 14.12. Renewable generation from biomass also contributes a substantial portion of the electricity generation in the S139 case, accounting for around 12% of total generation by end of the simulation. Growing dedicated crops (i.e., hybrid poplars) to fuel this generation would involve major land use changes in the western United States and Canada. Figure 14.14 shows that gas-fueled combined cycle units with carbon capture and sequestration would make a small contribution by the end of the simulation. This technology becomes available (at the high cost of 75$/MW) near the end of the simulation. In this scenario, carbon sequestration technology would provide around 2% of the generation at the end of the simulation. The appearance of this advanced “back-stop” technology marks the transition from the near term to the long term in the simulated system. Figure 14.15 summarizes the WECC results alongside of the EIA results for the country as a whole. The two studies show a similar reduction in CO2 emissions. S139 is projected to deliver a 75% reduction in CO2 emissions for the western United States and Canada (in the WECC study) and for the nation as a whole (in the EIA study). The retail rate impacts are projected to be smaller in the WECC study. The average retail electric rate is projected to be 23% higher by the year 2025 in the simulation with S139. The lower rate impacts are due to the different mix of generating resources and to the generally higher rates in the west. Figure 14.15 shows the results of additional simulations in the WECC study. The renewable energy production tax credit is an important incentive for wind and biomass generation. There is an “on again–off again” history on this particular incentive, so the simulations were repeated with the tax credit not in effect. Figure 14.15 shows that there would be less emissions reduction when the system was tested with the allowance prices shown in Fig. 14.8. With a carbon market, however, the smaller response of the electricity sector would probably lead to a somewhat higher price of allowances (since the electricity sector is the pivotal sector in bringing emissions into compliance with the cap). If the simulations were repeated with somewhat higher prices, the emissions reduction would be somewhat higher than shown in Fig. 14.15. The retail rate impacts would be somewhat higher as well. The remaining result in Fig. 14.15 shows the S139 impact in a scenario with more favorable assumptions for coal. The coal resources in the WECC are located in the eastern states and provinces, and there have been several proposals for major transmission projects to connect the eastern coal ﬁelds with the load centers on the coast. These projects were included in a new scenario that also included considerably higher prices for natural gas. And ﬁnally, the new scenario assumed slower growth in electricity demand, an assumption

EBL

7

This particular result would appear in many electric systems with the opportunity for fuel switching between coal and gas CCs. But it would not necessarily appear in a system like Australia which is dominated by coal. The impact of carbon prices in the Australian system is described in a separate chapter.

Competitive Electricity Markets

538

Percent reduction in CO2 emissions 100 Base case simulation for the western electricity system

The USA electricity sector (EIA)

75 Base case without the New production base case tax with credit slower growth and a shift to coal in the east

50

25

Industrial (EIA)

Transportation (EIA) Residential (EIA)

0 0

25

50

75

100

Percent increase in the retail price of energy Fig. 14.15. Comparison of simulated impact of S139 in the western electricity system with the EIA study of the US electricity sector (Ford, 2006).

which was more in line with recent industry forecasts. CO2 emissions would be cut by 63% and retail rates would be increased by 18%. However, if somewhat higher allowance prices were used in this simulation, the emission reduction would be somewhat higher. And the retail rate impact would be higher as well. The main conclusions from both the EIA study and the WECC study is evident from Fig. 14.15. The electricity sector results lie well above the 50/50 line used to distinguish between the responsive and unresponsive sectors of the economy. All the results show that the electricity sector would be highly responsive, which suggests that the electricity sector is likely to lead the way in the nation’s near-term response to carbon market legislation. An important insight from the WECC study is that the western electric system could deliver the major reductions in CO2 emissions by relying mainly on a price-induced demand reduction and through increased generation from existing technologies for renewable generation. The near-term results were possible without signiﬁcant reliance on advanced, carbon-free technologies. Both the EIA study and the WECC study deal with cap-and-trade, but the results also shed light on the impact of a carbon tax. For example, the EIA analysis provides an approximate indication of the impacts of a carbon tax that follows the upward trajectory shown in Fig. 14.8. (The results are approximate because the EIA analyses also included the market for offsets in S139.) The WECC analysis would yield the same result for both a carbon market and a carbon tax since the carbon prices in Fig. 14.8 were an input to the simulations. The differences between a carbon market and a carbon tax arise in an uncertain world where the impact of a particular tax trajectory cannot be estimated. With cap-and-trade, one hopes the market will ﬁnd the appropriate upward trajectory in prices (provided it is not prevented from doing so by a tight price-cap.) With a carbon tax, one hopes the government will remain committed to the upward trajectory for the tax over several decades. This commitment provides a predictable signal to investors, but it does not guarantee that emissions will be reduced sufﬁciently to meet a particular target.

EBL

Global Climate Change and the Electric Power Industry

539

14.13. Conclusions The IPCC scientists have reported that the evidence of global warming is unequivocal. They also report that anthropogenic emissions are very likely to have caused the warming. Policymakers around the world are calling for massive reductions in CO2 emissions to reduce the risks of global warming. Achieving these reductions will be a century-long challenge. Meeting this challenge in the long term will require continuation and expansion of research and development on carbon-free technologies such as carbon capture and sequestration. Meeting the challenge in the short term will require a mix of regulation and economic incentives to encourage investment in existing technologies. A carbon tax or a carbon market could provide the needed economic incentives, and there are good reasons for either policy. Policymakers in Europe and the United States are leaning toward carbon markets to provide the economic incentive. The ETS is nearing the end of a test period, and several cap-and-trade proposals are under consideration in the US Congress. Many believe the United States will soon adopt a mandatory carbon market at the federal level. Meanwhile, several states and regions are moving ahead with their own carbon reduction initiatives. This chapter explains the potential for the electric power industry to play a pivotal role in the coming years. The industry could lead the way in cutting CO2 emissions during the 10–20 years following the adoption of mandatory carbon markets. The reduced emissions will come largely from a reduction in conventional coal-ﬁred generation. The reduction in coal generation could be replaced by a combination of fuel-switching, demand reduction, and investment in renewable generating technologies that exist today.

EBL

References Bird, L., Holt, E., and Carroll, G. (2007). Implications of carbon regulation for green power markets. Technical Report NREL/TP-640-41076, National Renewable Energy Laboratory, April. Bowen, M. (2006). The messenger. Tech. Rev., 38–43. July/August. CARB (California Air Resources Board Market Advisory Committee). (2007). Recommendations for designing a greenhouse gas cap-and-trade system for California. Draft for public review, 1 June. Cambridge Energy Research Associates (2006). Survey of power industry executives. Restruct. Tod., 21, 1 September. Claussen, M., et al. (2002). Earth system models of intermediate complexity: closing the gap in the spectrum of climate system models. Clim. Dyn., 18, 579–86. Cooper, R. (2005). The Kyoto protocol: a ﬂawed concept. In Trade and Environment: Theory and Policy in the Context of EU Enlargement (Maxwell, J. and Reuveny, R., eds). Edward Elgar. Cooper, R. (2007). Comments on graduation and deepening. In Architectures for Agreement: Addressing Global Climate Change in the Post-Kyoto World (Aldy, J. and Stavins, R., eds). Cambridge University Press, Cambridge, UK. Cox, P.M., Betts, R.A., Jones, C.D., Spall, S.A., and Totterdell, I.J. (2000). Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model. Nature, 408(November), 184–7. CPUC (California Public Utilities Commission). (2007). PUC sets GHG emissions performance standard to help mitigate climate change. Docket # R.06-04-009, available at: www.cpuc.ca.gov. De Leyva, E. and Lekander, P. (2003). Climate change for Europe’s utilities. The McKin. Q., No. 1, 121–31. EC (2003). Directive 2003/87/EC establishing a scheme for greenhouse emission allowance trading within the community and amending Council Directive 96/61/EC. European Commission, Brussels. EIA (US Department of Energy, Energy Information Administration). (2003). Analysis of S139, the Climate Stewardship Act of 2003, June.

540

Competitive Electricity Markets

EIA (US Department of Energy, Energy Information Administration). (2006). International energy outlook 2006, June. Ellerman, A.D., Joskow, P.L., and Harrison, Jr., D. (2003). Emissions trading in the US: experience, lessons and considerations for greenhouse gases. Pew Center on Global Climate Change, May. Ellerman, A.D., Joskow, P., Schmalensee, R., Montero, J., and Bailey, E.M. (2000). Markets for Clean Air. Cambridge University Press. EPA (US Environmental Protection Agency). (2006). Acid Rain Program 2005 progress report. Report EPA-430-R-06-015, October. Fairﬁeld, H. (2007). When carbon is currency. The New York Times, 6 May. Fiddaman, T. (2002). Exploring policy options with a behavioral climate-economy model. Syst. Dyn. Rev., 18(2), 243–67. Fischer, C. and Morgenstern, R.D. (2006). Carbon abatement costs: why the wide range of estimates? Energy J., 27(2), 73–86. Flannery, T. (2005). The Weather Makers. Atlantic Monthly Press, New York, NY. Ford, A. (1999). Modeling the Environment. Island Press. Washington, DC. Ford, A. (2006). Simulating the impact of a carbon market on the electricity system in the western US. Proceedings of the 24th International Conference of the System Dynamics Society, Nijmegen, The Netherlands, July. Forest, C., Webster, M., and Reilly, J. (2004). Narrowing uncertainty in global climate change. The Ind. Phys., August/September, 20–5. Fouquet, D., Grotz, C., Swain, J., and Vassilakos, N. (2005). Reﬂections on a Possible Uniﬁed EU Financial Support System for Renewable Energy Systems. European Renewable Energies Federation and the Worldwatch Institute, January. Gill, R.A., Anderson, L.J., Polley, H.W., Johnson, H.B., and Jackson R.B. (2006). Potential nitrogen constraints on soil carbon sequestration under low and elevated atmospheric CO2 . Ecology, 87(1), 41–52. Gore, A. (2006). An Inconvenient Truth. Rodale, Emmaus, PA. Govindasamy, B., Thompson, S., Mirin A., Wickett, M., Caldeira, K., and Delire, C. (2005). Increase of carbon cycle feedback with climate sensitivity. Tellus, 57B, 153–63. Hansen, J., Sato, M., Kharecha, P., Russell, G., Lea, D., and Siddall, M. (2007). Climate change and trace gases. Philos. Trans. R. Soc., A, 365, 1925–54. Harmon, R. and Hirschhorn, M. (2006). Cleaning the air: the impact of carbon cap and trade regulations on renewable energy. SOLAR 2006 Conference Proceedings, July. Houghton J. (2004). Global Warming: The Complete Brieﬁng, 3rd edn. Cambridge University Press. IPCC (1997). An introduction to simple climate models used in the IPCC second assessment report. IPCC (2001). Climate Change 2001: The Scientiﬁc Basis, Summary for Policymakers, available at: www.ipcc.ch/. IPCC (2005). Special report on carbon dioxide capture and storage. IPCC (2007). Climate Change 2007: The Physical Science Basis, Summary for Policymakers, available at: www.ipcc.ch/. Jacoby, H.D. (2007). Interpreting the Stern Review on the economics of climate change. Testimony to the Energy Committee, US Senate, 13 February. Jacoby, H.D. and Ellerman, A.D. (2004). The safety valve and climate policy. Enerygy Policy, 32, 481–91. Jones, C.D., Cox, P., and Huntingford, C. (2003). Uncertainty in climate-carbon-cycle projections associated with sensitivity of soil respiration to temperature. Tellus, 55B, 642–8. Kerr, R. (2007). Pushing the scary side of global warming. Science, 316 (8 June), 1412–5. Kump, L.R. (2002). Reducing uncertainty about carbon dioxide as a climate driver. Nature, 419 (12 September), 188–90. Lohr, S. (2006). The cost of an overheated planet. The New York Times, 12 December. Lomborg, B. (2006). Stern Review. The Wall Street Journal, 2 November, available at: http://www. wsj.com. Luoma, J. (1991). Gazing into our greenhouse future. Audubon, March. Mastrandrea, M.D. and Schneider, S.H. (2004). Probabilistic integrated assessment of “dangerous” climate change. Science, 304, 571–5, (23 April).

EBL

Global Climate Change and the Electric Power Industry

541

Mendelsohn, R.O. (2006). A critique of the Stern report. Regulation (a magazine of the CATO Institute), Winter 2006–2007, available at: http://www.cato.org/pubs/regulation/index.html. Meyer, A. (2000). Contraction and Convergence: The Global Solution to Climate Change, Schumacher Brieﬁng No 5. Green Books. MIT (Palstev, S., Reilly, J., Jacoby, H., Ellerman, A.D., and Tay, K.) (2003). Emissions trading to reduce greenhouse gas emissions in the United States: The McCain–Lieberman Proposal. Report No. 97, MIT Joint Program on the Science and Policy of Global Change, 17 June. MIT (2006). Website information on the integrated global systems model, available at: http://web.mit. edu/globalchange/www/if.html. MIT (2007a). The future of coal: options for a carbon-constrained world, available at: http://web.mit. edu/coal/. MIT (Paltsev, S., Reilly, J., Jacoby, H., Gurgel, A., Metcalf, G., Sokolov, A., and Holak, J.) (2007b). Assessment of US cap-and-trade proposals. Report in progress at the MIT Joint Program on the Science and Policy of Global Change, available at: http://web.mit.edu/globalchange/www/. NAS (National Academy of Sciences, Committee on Abrupt Climate Change) (2002). Abrupt Climate Change: Inevitable Surprises. National Academy Press, Washington DC. NASA (National Aeronautical and Space Administration) (2006a). Website feature on “2005 warmest year in over a century,” available at: http://www.nasa.gov/vision/earth/ environment/2005_warmest.html. NASA (National Aeronautical and Space Administration) (2006b). NASA Earth Observatory website, available at: http://earthobservatory.nasa.gov/Library/CarbonCycle/carbon_cycle4.html. NECP (National Commission on Energy Policy) (2004). Ending the energy stalemate, December. NECP (National Commission on Energy Policy) (2006). Design issues for market-based greenhouse gas reduction strategies workshop series, Cambridge Energy Research Associates, available at: www.energycommission.org. NECP (National Commission on Energy Policy) (2007). Energy policy recommendations to the President and the 110th Congress, April. NRDC (Natural Resources Defense Council) (2006). How low should we go? A legislative guide for reducing global warming pollution, available at: www.nrdc.org/policy. Oregon Strategy (Oregon Governor’s Advisory Group) (2004). Oregon strategy for greenhouse gas reductions, available at: http://www.oregon.gov/ENERGY/GBLWRM/Strategy.shtml, accessed December 2004. Pew Center (2005). The European Union emissions trading scheme (EU-ETS): insights and opportunities. A Report, Pew Center, Alexandria, VA. Pizer, W. (1999). Choosing price or quantity controls for greenhouse gases, climate issues. Brief No. 17, Resources for the Future, Washington, DC, July. Portney, P. (2005). A carbon tax in time, saves nine. Inside EPA, 23 June. President G.W. Bush (2007). Twenty in ten: strengthening America’s energy security. 2007 State of the Union Policy Initiatives in Focus: Energy. Prinn, R., et al. (1999). Integrated global system model for climate policy assessment: feedbacks and sensitivity studies. Clim. Change, 41, 469–546. Reilly, J., et al. (1999). Multi-gas assessment of the Kyoto Protocol. Nature, 401, 549–55. Reilly, J. and Paltsev, S. (2005). An analysis of the European emission trading scheme. Report No. 127, MIT Joint Program on the Science and Policy of Global Change, October. RGGI (Regional Greenhouse Gas Initiative) (2006). An initiative of the Northeast and Mid-Atlantic states of the US, available at: http://www.rggi.org/index.htm. Samuelsohn, D. (2007). Californians’ trip raises prospect of EU-US carbon market. Greenwire, E & E publishing service, 5 April. Sarmiento, J.L., Quere, C.Le, and Pacala, S.W. (1995). Limiting future atmospheric carbon dioxide. Glob. Geochem. Cycles, 9(1), 121–37. Schneider, S.H. (1990). Global Warming. Vintage Books, New York, NY. Socolow, R., Hotinski, R., Greenblatt, J., and Pacala, S. (2004). Solving the climate problem. Environment, December. Sterman, J.D. and Sweeney, L.B. (2007). Understanding public complacency about climate change. Clim. Change, 80(3–4), February.

EBL

542

Competitive Electricity Markets

Stern Review (H.M. Treasury) (2006). Stern Review on the Economics of Climate Change. Cambridge University Press. Also available from the Independent Reviews section of the UK HM Treasury website http://www.hm-treasury.gov.uk/. Sterzinger, G. (2006). Transforming production tax credits. Pub. Util. Fort., July. Technology Review (2006). Massachusetts Institute of Technology, Cambridge, MA. Tellus (Bailie, A., Dougherty, B., Heaps, C., and Lazarus, M.) (2004). Turning the corner on global warming emissions: an analysis of ten strategies for California, Oregon and Washington. Tellus Institute draft report for the West Coast Governor’s Global Warming Initiative, 28 July. The Economist (2006). Special Survey, The heat is on, 9 September. The Royal Society (2005). Ocean acidiﬁcation due to increasing atmospheric carbon dioxide. Policy document 12/05, www.royalsoc.ac.uk, accessed June 2005. Tol, R. (2006). The Stern Review of the economics of climate change: a comment. Institute for Environmental Studies, Vrije Universiteit, Amsterdam, The Netherlands, 2 November. UK Energy and the Environment (2006). Press Release, 18 September. UNFCCC (1992). United Nations framework convention on climate change, available at www. unfccc.int. UNEP (2006). United Nations environmental program, available at: http://www.unep.org/. Varian, H. (2006). Recalculating the costs of global climate change. The New York Times, 14 December. Weart, S.R. (2003). The Discovery of Global Warming: New Histories of Science, Technology and Medicine. Harvard University Press, Cambridge, MA. Weart, S.R. (2007). The discovery of rapid climate change, available at: Physics Today.org, April. Webster, M., et al. (2003). Uncertainty analysis of climate change and policy response. Clim. Change, 61, 295–320. Weyant, J. and Hill, J. (1999). The costs of the Kyoto Protocol: a multi-model evaluation. Energy J., 1–390. World Bank (Cooper, K. and Ambrosi, P.) (2006). State and Trends of the Carbon Market 2006. The World Bank, Washington, DC, May. WRCAI (Western Regional Climate Action Initiative) (2007). Western Regional Climate Action Initiative, signed by the Governors of Washington, Oregon, California, Arizona, and New Mexico, 26 February.

EBL

Chapter 15 Reform of the Reforms in Brazil: Problems and Solutions JOÃO LIZARDO R. HERMES DE ARAÚJO,1 AGNES MARIA DE ARAGÃO DA COSTA,2 TIAGO CORREIA,2 AND ELBIA MELO3 Centre for Electric Energy Research, Rio de Janeiro, Brazil; 2 Brazilian Ministry of Mines and Energy, Brasília, Brazil; 3 Chamber of Electric Energy Commercialization (CCEE), São Paulo, Brazil

1

Summary

EBL

In the Brazilian electricity supply industry (ESI), the ﬁrst market-oriented reform ran into trouble, leading to insufﬁcient investment in generation and higher consumer tariffs. The reform of the reform, implemented by the Lula administration, purports to overcome problems through a combination of competitive bilateral contracts between consumers, traders, distributors, and producers, and regulated contracts between generators and distributors through auctions. This chapter analyzes the characteristics of both reforms and the experience so far with the new setup. 15.1. Introduction: Reform as a Learning Process Restructuring an industry is not an easy task. When that industry has long been universally considered as a natural monopoly, and operated as a single vertically integrated regulated monopoly, or as a coordination of local integrated monopolies, to turn it into a competitive industry, even when feasible, requires great care and attention to detail. When, moreover, that industry has peculiarities that require real-time coordination of physical operation to guarantee system stability, and shows large maturation lags and asset speciﬁcity for investment, the task is really daunting. More than twenty years ago, Joskow and Schmalensee (1983) had already discussed the difﬁculties awaiting would-be electric power reformers. Nevertheless, after the Chilean and British experiments, marketoriented reforms of electricity supply industries gradually spread throughout the world. It is not surprising that many of these reforms have required adjustments and, in some cases, substantial changes, as further described in the Introduction to this volume by Sioshansi. Clearly, electricity sector reform is a learning process; and while some general inferences 543

544

Competitive Electricity Markets

may be drawn, enough peculiarities remain to make each case special, and challenging. Two examples of successful reforms may be in order to emphasize the latter point. The forerunner of all, the Chilean reform (Cf. Raineri, 2006) turned around two focal points: competition in expansion (with regulated marginal cost pricing for system prices, centralized operation, and regulated distribution), and free contracts for large consumers. It faced three crises, which led to adjustments: a drought; too vigorous growth of installed capacity in the North leading to system imbalances; and gas supply constraints by Argentina. The ﬁrst and the third crises may be taken as basically exogenous to reform, belonging rather in the realm of broader energy policy; but the second crisis appears to have been endogenous – the success of the setup in attracting investment was excessive and required measures to ensure better coordination of electricity industry players. By all accounts, the Chilean reform has responded well and may be considered successful. The British reform has served as a model for many other experiments, and may be considered as one of the best (see account by Newbery, 2006). Differently from Chile, it is focused on market bids rather than marginal cost. But like Chile, it also went through several reforms. The initial pool model was handicapped by the effective duopoly of National Power and PowerGen, which required active and repeated interventions by the regulator; in 1998, it was replaced by NETA (New Electricity Trading Arrangement) that differed from the Pool in several signiﬁcant aspects – it was more akin to the Nord Pool scheme and did away with capacity charges, one of the ways in which the duopoly used their market power through capacity bids. Several critics (including Newbery) questioned the need for change at a time when market structure was becoming less concentrated. The creation of NETA went with a change in regulator – OFFER (Ofﬁce of Electricity Regulation) and OFGAS (Ofﬁce of Gas Regulation) were merged into OFGEM (Ofﬁce of Gas and Electricity Markets), which acquired a committee structure rather than centering on the Director-General, a broadening of scope to include Scotland, and a push for retail competition. In the end, the balance appears favorable to the British reform – or reforms. The Brazilian reform has also gone through at least three phases: an initial scramble when privatization and reform followed nearly independent paths between 1995 and 2000, an attempt at mending things from 2000 through 2003, and a more substantial reform in 2004. The process is discussed at length in Araújo (2006), and we present here only a brief outline, highlighting the main issues. Reform started by selling distribution assets, before a regulatory framework had been established. This led to considerable regulatory backlogs, not least of which was a sharp rise in ﬁnal electricity prices after the 1999 Real devaluation; to this day, consumer electricity prices are much above those of countries with a similar generation structure like Canada, because of contractual clauses. This drained away popular support for the reform. Another critical problem of the ﬁrst reform was the matter of contracts and investment in generation. The ﬁrst reform was molded as an adaptation of the British Pool model to a hydro-dominated system, with bids by thermal generators but not by hydropower plants; the pool price was calculated as a marginal cost like in Chile, the difference being that in Brazil the complexity of the system ruled out dealing with hydropower plants as a single plant, and transmission constraints are also far more complex. Contracts were initially assigned to generators and distribution companies, to be gradually phased out, and it was hoped that spontaneous bilateral contracts would develop from the working of the spot market – this never happened, and was one of the main causes for the 2001 crisis. The spot market also had governance troubles that haunted its operation through the ﬁrst and the second phases. The 2004 reform made a complete overhaul of governance as well as of investment and contracting schemes.

EBL

Reform of the Reforms in Brazil

545

Despite many speciﬁc aspects of the Brazilian ESI, the learning process in its reform may be used to draw parallels elsewhere; this chapter uses the case of Brazilian reform to discuss this process, and will draw lessons from other chapters of the book where relevant. The reader may also want to read Chapter 2 by Correljé and De Vries in this volume and identify relevant patterns. Section 15.2 summarizes the main features of the ﬁrst reform, pointing out successes and failures as it developed; Section 15.3 discusses the investment and contracting issues in electricity reforms; Section 15.4 introduces the second reform, and its contracting framework; Section 15.5 presents and discusses the Brazilian experience with energy auctions, a cornerstone of the new framework; ﬁnally, Section 15.6 provides the conclusions with a brief assessment of the new set-up and the issues still remaining. 15.2. The First Reform: Features, Successes, Failures 15.2.1. The ﬁrst phase, 1995–2000 Brazil is a large country (8.5 million square kilometers, larger than the USA without Alaska), with continental distances, a population of 186 million in 2006, per capita GDP US$4300 at current exchange rates in 2005 and a sizable power system, both in terms of generating capacity (Table 15.1) and of grid extension, with 83 000 kilometers of transmission lines at 230 kV and above (Fig. 15.1). A prominent feature of the Brazilian power system is the weight of hydropower, like Canada and Norway, but with very little international trade, unlike either of those. The Brazilian hydropower system was built upon big plants, with large reservoirs. Of the 144 plants with more than 30 MW of installed capacity, 21 have reservoirs with more than 5 million m3 and 32 have reservoirs with capacity of regulation over several years. The signiﬁcant hydropotential, both operational and to be developed, has shaped Brazilian power policy for the better part of a century. It has also played a prominent role in both the power sector reforms undertaken in the last decade, led respectively by the Cardoso and the Lula administrations. As discussed in Araújo (2006), the Brazilian reform effectively started in 1995, with three distinct sets of measures: First, several pieces of legislation were enacted – altering the concession regime, forcing utilities to ﬁnish projects or giving up concessions, and mandating open access for large consumers and independent power producers – which stimulated consortia for investing in power plants. At the same time, a study to restructure

EBL

Table 15.1. Generating capacity and generation by source, and imports

Conventional – Hydro Conventional – Thermal Nuclear Other TOTAL

Operational and monitored capacity (GW) *

Generation in the interconnected system (TWh) †

738 203 20 20 981 GW

382 20 14 02 4162 TWh

Net international imports not counting Itaipu ‡ NA NA NA NA 0.5 TWh

* Data for May 4, 2007; ANEEL, “Banco de Informações de Geração”. † Data for January to December 2006; ONS, “Histórico da Geração de Energia” (includes Itaipu). ‡ Data for January to December 2006; ONS, “Histórico da Operação”.

546

Competitive Electricity Markets

EBL

Fig. 15.1. The Brazilian transmission system (138 kV and higher). Source: ONS, Sistema de Transmissão Horizonte 2007.

the power sector was launched, aiming to introduce competition and to divest all of distribution, transmission, and generation, excepting nuclear plants and the Brazilian half of Itaipu1 . Also in the same year, divestment of federal assets began – before any restructuring and regulatory framework had been put in place. The electricity regulator ANEEL (Agência Nacional de Energia El´ectrica) was set up and started work only in December 1997, by which time 10 distribution utilities had been divested for US$ 12 billion. The reform proceeded in 1998 with the creation of ONS (National System Operator), which took over the functions of an existing organism GCOI 1 Itaipu was jointly built by Brazil and Paraguay at the frontier of both countries. The 1973 Itaipu Treaty established that all the energy produced by Itaipu and not used by Paraguay must be purchased by the distribution utilities of the Brazilian Southern, Southeastern, and Center-Western regions; Itaipu energy prices are quoted in US dollars.

Reform of the Reforms in Brazil

547

(Grupo de Corrdenção da Operação Interligada) to operate the physical system, and of a Bulk Market Organism (MAE), which was a real novelty and did not function smoothly. Besides basic problems to be discussed below, the MAE implementation also ran into hurdles. As described in Araújo (2006), MAE was set up as a self-regulating body of market agents, following the Californian model. Without due attention to conﬂict-resolving mechanisms, the result was that many important points remained undeﬁned in the market rules. With ﬂawed governance, large actors (mostly state-owned ﬁrms) disregarded rules and helped bring the bulk market to a crisis. Overall, this ﬁrst phase was characterized by a misalignment between restructuring, regulation, and divestment, which proceeded in parallel rather than sequentially. This had important consequences: First and foremost, it increased risks to investors, requiring inefﬁcient clauses in concession contracts to attract buyers. This substantially increased the difﬁculty of regulation, and ultimately led to several failures and to an incomplete reform. Another consequence was that expectations on the reform schedule were unduly sanguine, and led to cuts in investment by federal generators under assumption that private investment would replace them; the lack of materialization of this led to the 2001 crisis.

15.2.2. The second phase, 2000–2003 This phase was marked by correction and adjustments to the reform mechanisms, and coping with the 2001 crisis and rationing. Since the latter required a set of emergency measures, there were a number of conﬂicts between both purposes. The result was a patchwork setup that enhanced the need for an in-depth revision. MAE started operating in September 2000; accounting should be done every month, but the ﬁrst balance sheet only appeared in the second half of 2002. Even worse, ﬁnancial settlement only started being made – partially – in January 2003. This happened for a host of factors, mainly conﬂicts among market agents; but there were also problems with the regulator and with the market rules themselves. The ﬁrst paralyzing conﬂict happened between a generator and distribution utilities. To start market operations, ANEEL had assigned initial contracts between the generators and the distribution companies (these were to be progressively discontinued from 2002 onwards). According to MAE rules, a generator had to deliver the amount of energy contracted with distribution companies, whether it had generated energy or not. If necessary, the generator had to buy energy in the spot market to honor the contract. Initial contracts had been issued for the energy to be generated by the Angra II nuclear plant, when it was still under construction. A four-month delay in the date when Angra II started operation made its contracts exposed to spot market prices during this time (which were then very low). Distribution utilities demanded that this energy be sold at spot prices, while the generator alleged that nuclear plants should not be commercialized at MAE, due to a special treatment granted by law 9648/98 (which was true, also for the binational Itaipu plant). The consequences of the episode were very serious for the sector. First, divergence as to contractual engagement of the Angra II plant paralyzed the bulk market only three months after it started operation, remaining so until January 2002, without settlement. This resulted in loss of credibility for MAE agents, jeopardizing competition. Another consequence was that agents themselves perceived that engagements could be broken; this raised the specter of moral risk, further degrading market performance.

EBL

548

Competitive Electricity Markets

This affair called the attention of the regulator to the need for monitoring the behavior of agents in the Bulk Power Market, and in 2001, ANEEL introduced changes in rules and in governance, with a revision of the market agreement to include mechanisms for guarantees and penalties (nonexistent till then). However, until August 2001 the new agreement had not been signed because of boycotting by agents. These impasses also impeded regular accounting and settlement, and other problems appeared, like the case of Annex V of Initial Contracts in the context of the rationing that started in June 20012 . This was similar to the Angra II affair and caused large losses to generators. The problems detected led the regulator to intervene a second time, appointing new councillors to MAE and restructuring again its governance mechanisms to avoid disproportionate decision power by vertically integrated agents. MAE was thus reorganized under special legal status and with a speciﬁc management regime through Law 10433 of 24 April 2002. By this law, MAE became a private notfor-proﬁt entity, subject to ANEEL authorization, regulation, and monitoring. Its General Assembly was also changed: vote by categories was abolished, meaning that one enterprise could vote only once (in the earlier model, an integrated company could vote both as generator and as distributor). Despite the reforms made, the Bulk Market only started partial settlement in January 2003, as even after the changes made by Law 10.433/2002 there remained a few conﬂicts and disputes among agents. The 2001 crisis, on the other hand, required measures ranging from rationing to special contracts for merchant plants and a number of exceptions to regulation, to encourage new investment in generation capacity. A number of ad hoc contracts were signed, leading to fresh conﬂicts after the rationing ended.

EBL

15.2.3. Main features, successes, and failures of the ﬁrst reform The ﬁrst reform had most of the ingredients postulated in the literature: a regulator, an independent system operator, a bulk market operator, open access, a spot market, bilateral contracting, regulation of the wires business, operational if not factual unbundling of generation, transmission, distribution, and trading. What worked, what did not, and why? To answer these questions, it is best to look separately at distribution, transmission, and generation. As mentioned above, distribution utilities were the ﬁrst to be divested, and most of them are now investor-owned, responding for 66% of the revenues generated in distribution activity in 2003. In the Brazilian Privatization Program (PND), concession contracts were awarded to the bidder with the highest offer, taking existing tariffs as given and subjected to a price-cap scheme with a readjustment index linked to commodity prices. The low market risk of the segment made them attractive and easy to sell, even after exceedingly generous clauses (e.g., quality of service requirements, or a zero X-factor for the ﬁrst eight years) common in the ﬁrst concession contracts were dropped or altered after serious failures in 1998. After the ﬁrst teething troubles, the segment appeared set to start working smoothly when disaster struck in the form of the 1999 Real devaluation. The readjustment index offset the devaluation losses only partially, and investors found that their starting assumptions no longer held true. Some utilities had even contracted debts with their overseas owners, and as a result found themselves in ﬁnancial straits. On the other hand, consumers found that their tariffs were rising above inﬂation, contrary to what they 2

See Araújo (2006) for an account of the 2001 crisis and rationing.

Reform of the Reforms in Brazil

549

had been promised, and grumbled against the terms of concessions. This state of affairs persisted from 1999 to 2001, complicated by troubles in the bulk market (to be discussed below). The 2001 rationing dealt another serious blow to distribution utilities since consumption fell 25% and self-production increased; utilities asked for compensations, and got them. First, tariffs were increased to make up for lost sales; second, restrictions on verticalization were softened, as were restrictions on cost pass-through to consumers, in order to stimulate addition of new generating capacity to the system as quickly as possible. Despite all these measures, some utilities could not become solvent and either sold out to third parties or returned the concession. These latter were for the most part utilities in poorer states in the North and Northeast; state governments were unable or unwilling to take on the burden, and they were taken on temporarily by the Eletrobrás Group to sanitize before a resale. To consumers, the net result was another raise in tariffs, which led to more public disaffection with the reform. In contrast, with transmission things went very smoothly. Since the wires business was agreed by all to be a natural monopoly, the procedures adopted were very different. Although reformers toyed with the idea of setting up one or two dedicated Transmission Companies, in the end it was decided to split ownership and operation. Existing assets would receive a revenue (initially ﬁxed by the regulator), yearly readjusted according to an index, following a price-cap scheme. The National System Operator (ONS) would make plans for grid expansion; planned additions would be auctioned, going to the bidder accepting the lowest revenue for the line under auction (yearly readjusted according to the same index). In all the cases, asset owners are required to maintain and operate their asset as ordered by ONS. This simple scheme has worked quite well, with lively competition among national and international agents. Generation was the segment that fared worst in the reform, in several aspects. It was very hard to divest, for reasons examined elsewhere (Araújo, 2001, 2006). Its working was closely linked to the bulk market, since contracts would be freely signed after an initial, adaptation period. Accordingly, federal generators were to be prepared for divestment and curtailed in their investment plans. As to investment in generation, the ﬁrst reform hoped that liberalization would sufﬁce for that. This is understandable, in view of the lack of reform experience prevailing at the time. Some concern with the weight of hydropower was shown by a few measures, which unfortunately did not work as expected. First, thermal plants could choose to declare part or all of their energy to be inﬂexible, meaning that it must be dispatched regardless of its merit order. This might have worked as inducement to investment, if this energy were paid according to its cost or price bid; but it was felt that this might bring serious distortions such as incentives toward inefﬁcient choices, and might shut down cheap hydropower plants when water was abundant. Inﬂexible energy thus received system price; as will be shown in Section 15.3, this is normally far below the operating cost of thermal plants. Consequently, the inﬂexibility clause became void of meaning. In 1999, when it became clear that thermal plants were not coming on as expected, the Federal Government decided to launch a Priority Program for Thermal Plants (PPT), with some favored clauses, aiming to build 54 thermal plants until 2002, for a total 19 GW capacity. As it turned out, when the 2001 crisis broke out, only a dozen gas-ﬁred plants were under construction, totalling around 3 GW, most of which had Petrobrás as a partner or were self-generation projects by industrialists concerned with the security of their supply; other investors were more than reticent. Investment in hydropower was also hindered by several factors, the most signiﬁcant of which were the risks involved in long-term

EBL

550

Competitive Electricity Markets

maturation investments, in a market environment with rules as yet in the making, and the fact that most large investors were reserving themselves for buying off generators rather than for expansion projects. A more comprehensive treatment of the investment issue and of the 2001 crisis may be found in Araújo (2001, 2006) and needs no further development here. It remains to brieﬂy mention the problems of the bulk market, to be more extensively discussed below. The MAE problems had no direct connection with the 2001 rationing, although the credibility loss they entailed was of consequence; MAE was blocked almost from the start, and throughout the crisis it remained effectively paralyzed. One might say that, even had MAE operated normally, the crisis would have occurred for the causes pointed in Section 15.3 and in the paragraph above: the (working) spot market would not be an adequate signaller for ﬁnancial instruments and for long-term contracts in view of the peculiarities of the Brazilian system, and long-term contracts would not develop easily by themselves. On the other hand, the MAE ills arose from a badly implemented, ﬂawed conception of market governance and market rules. In either case, there were design ﬂaws and underestimation of crucial details.

15.3. The Investment and Contracting Issue One outstanding consequence of market-oriented reforms in the electricity supply industry worldwide has been a change in the investment context. Risks that had formerly been borne by consumers – or by taxpayers – have now to be faced by investors, with widely varying results, as witness an assessment by Paul Joskow: “The jury is still out on whether and how competitive power markets can stimulate appropriate levels of investment in new generating capacity in the right places at the right times.”3 This section discusses one approach to this issue in the context of Brazil, central to the reform of the reform. Other approaches – to distinct systems – are discussed in other chapters of this book, notably by Adib et al. for the USA , by Moran and Skinner for Australia in Chapter 11, and a systematic discussion is made by Correljé and De Vries in Chapter 2). First of all, one must bear in mind the fact that the power industry in Brazil is highly dependent on the water storage capacity of the reservoirs located in many regions of the country and on the transmission capacity between regions, as roughly 85% of the power generated in Brazil, and three-fourths of installed capacity, is hydroelectric. The historical dependence of Brazil on hydropower is clearly shown in Fig. 15.2. Almost 98% of the Brazilian power market is covered by an interconnected transmission network designed to enable power interchange among the Northeastern, the Southern, and the Southeastern/Mid-Western regions of the country (see Fig. 15.1). Moreover, hydropower plants are spread over a dozen major river basins, with marked differences in hydrological regimes. Thermal generation is being expanded in order to improve the security of power supply, since reservoir level depends not only on the hydrology but also on the other uses of water; and new hydropower plants have signiﬁcantly smaller reservoirs, with less ﬁrm energy relative to nameplate capacity. This will require signiﬁcant investment in both thermal and hydropower plants. Secondly, let us brieﬂy recall some issues that have been raised by electricity reforms concerning uncertainty, volatility, and investment in power generation. Putting it bluntly,

EBL

3

See Joskow (2006) p. 20.

Reform of the Reforms in Brazil

551

100% 90% 80% 70% 60% Nuclear% Thermal% Hydro%

50% 40% 30% 20% 10%

19 20 19 32 19 35 19 38 19 41 19 44 19 47 19 50 19 53 19 56 19 59 19 62 19 65 19 68 19 71 19 74 19 77 19 80 19 83 19 86 19 89 19 92 19 95 19 98 20 01

0%

Year Fig. 15.2. Installed capacity share, 1920–2003. Source: Araújo (2006), Araújo and Besnosik (1992).

risks associated with new investment have been transferred from consumers to investors. Moreover, while regulatory risks are decreasing – both from the change to a marketbased context and from learning by regulators and by ﬁrms through experience – other risks, mostly, but not exclusively, market-related, have claimed attention in recent years. These broadly fall into two categories: risks regarding fuel prices or electricity prices, and licensing risks. In order to keep the argument simple, only the ﬁrst category shall be discussed, although licensing risks now play a large role in Brazil, and are expected to grow as environmental issues become more important. Fuel and electricity price uncertainty affects different generation technologies in distinct ways, according to their capital composition: capital-intensive technologies are more sensitive to electricity price uncertainty and fuel-intensive technologies to fuel price uncertainty (IEA, 2003). However, this is not the whole story: electricity price uncertainty may seriously affect investment in gas-ﬁred plants in a hydro-dominated context such as that in Brazil (Araújo, 2006). This uncertainty is accompanied by substantial volatility4 in power markets and must be hedged against for investment to take place. Financial hedges have been used in several experiments, most notably in the Nord Pool, but to some extent also in England & Wales (E&W), in Germany, and in the USA with some success. In the E&W case, ﬁnancial hedging mechanisms took the form of “Contracts for Differences” (CFD) during the ﬁrst (“the Pool”) reform phase, and were thus partly contract hedges; the NETA reform created no speciﬁc ﬁnancial hedges, although deals in futures/forward contracts have grown in importance over the years (IEA, 2003). In the Nord Pool, speciﬁc markets were created for this purpose (Eltermin, Eloption), and a variety of alternatives exist for forward/futures contracts, with active Over the Counter

EBL

4

Volatility should not be confused with uncertainty, being rather a measure of the size and suddenness of changes. One could deﬁne volatility as the rate of change in uncertainty (as in IEA, 2003), but in the limit one could have volatility without uncertainty. Whatever the precise deﬁnition, power markets present considerable volatility as an analysis of spot prices in Nord Pool, PJM, or ERCOT will show.

Competitive Electricity Markets

552

(OTC) dealing in these instruments5 . Elsewhere, ﬁnancial hedging has been less used; and in the USA there has even been a decrease following the Enron debacle. The point is that, contrary to commodity markets, ﬁnancial power markets are not liquid enough for ﬁnancial hedges to sufﬁce. Even in E&W and in Nord Pool, the ratio of traded amount to physical demand (a surrogate for liquidity, however imperfect) is only 8 to 9 times, while in the German EEX it is 2.5 and in France a mere 0.005; in comparison, the accepted benchmark for a commodity market to be considered liquid is 25 times 6 (IEA, 2003). In other words, at least till now power markets do not behave like liquid commodity markets (we do not enter into the issue of whether power may be considered a bona ﬁde commodity). In the Brazilian case, another complication arises from the fact that overall ﬁnancial markets are much less liquid than more mature markets, by one or two orders of magnitude. As an illustration (cf. Fig. 15.3), the ratio of daily volume of transactions in OTC derivatives to estimated daily GDP in April 2001 was 0.88 in Brazil, very small7 compared to the USA (where the ratio was 7.6 at the same time), Asia (10.96), or Europe (31.82 overall; in the UK, the ratio was as high as 108.42) 8 . Added to the observations above, this means that ﬁnancial hedges are out of the question in Brazil, as mechanisms to attract investment in generation expansion. For this reason, early on, contract hedges were chosen in the Brazilian reform as the preferred hedging mechanism to deal with investment risks. This was not an easy task, and ﬁrst attempts failed as discussed elsewhere9 . The preceding section described some

EBL

Ratio between daily OTC transactions and daily GDP in selected markets 120 100 80 60 40 20 0 Brazil

USA

Asia

European Union

UK

Fig. 15.3. Stock market liquidity. Source: Fernandez (2002).

5

Amundsen et al. (2006). For a detailed look at Nord Pool ﬁnancial instruments, see also Benth and Koekebakker (2005). 6 International Energy Agency (2003). 7 There is not space to go into the causes for this; one important factor, however, is the weight of instruments relating to public debt with high real interest rates, which dwarf others. 8 Cf. Table 12 in Fernandez (2002). 9 Araújo (2001), see also Araújo (2006).

Reform of the Reforms in Brazil

553

failings of the initial reform in this respect. Here, however, the discussion centers on the underlying cause for those failures, drawing from previous work by one of the authors. The large share of hydropower in electricity supply is another issue. Other countries, like Canada and Norway, also have a large share of hydropower generation; but Brazil has very little electricity trade with neighboring thermal-based systems, unlike either of those10 . This has important consequences, which were not duly appreciated in the early reform years. Basically, hydropower has high capital costs and very low operating costs. Thus, when water is abundant in a hydro-based system, the short-term electricity marginal cost is very low; on the other hand, when water is scarce the cost will be that of the backstop generation technology; prices may go much higher if supply cannot meet demand in the short run. These are well-known facts, and Nord Pool prices show very large seasonal variations, as well as from year to year (notably in 2002), because of the sizeable hydropower share in the Nordic market (around 51% both of installed capacity and of effective generation)11 . The difference with Brazil lies in two main aspects: the much higher share of hydropower in Brazil and the importance of existing reservoirs as energy regulators. The very large share of hydropower in the Brazilian power plant park means that most (80–90%) of the time hydropower plants are the marginal producer. The fact that most hydrocapacity in Brazil is associated with large reservoirs (which used to have the ability for multiyear regulation – ﬁrm energy was traditionally calculated with respect to a critical ﬁve-year period12 ) means that years may go by without a thermal plant being dispatched in the interconnected system. This is in stark contrast to the Nordic situation, where half of the time thermal plants are in operation. As shown in Araújo (2001), lack of investment since the 1980s has reduced the capacity for regulation of Brazilian reservoirs, and new planned hydropower plants will have smaller reservoirs to reduce environmental impacts (which also reduces the regulation period). Also, the determination of spot prices in the Brazilian system embeds water inﬂow forecasts (and thus, future thermal generation) in the estimation of system marginal cost (SMC) (Araújo, 2006), so that SMC rises above operating costs of hydropower plants when reservoirs are low. These two factors have caused spot prices to oscillate between roughly US$50/MWh and US$70/MWh in the second half of 2000, when the bulk market started operation and supply was rather tight. Nevertheless, as Fig. 15.4 shows, since the 2001 rationing ended spot prices for heavy loads in the Southeast (two-thirds of the Brazilian market) remained near US$6/MWh most of the time, and below US$18/MWh all the time13 ). In other words, in “normal” years the reservoirs are still able to smooth out seasonal inﬂow variations and thus rule out thermal dispatching save in exceptional circumstances. This would only change if the share of thermal power were large enough to make thermal plants the marginal producer for a signiﬁcant fraction of the time.

EBL

10

The Norwegian market is so integrated in the Nordic market that it can hardly be analyzed outside it. 11 Cf. Amundsen et al. (2006). 12 This has now been replaced by a stochastic version: the probability of an energy deﬁcit is less than 5% in any given year; that is, one in twenty years may run a deﬁcit (of any size). 13 All ﬁgures are in current US dollars. The exchange rate varied a lot during this period: from around 2 R$/US$ in 2000, it went up to a new level around 2.5 from mid-2001 till mid-2002, when election fright took it to roughly 4; from mid-2003 to mid-2004 it remained in the neighborhood of 3 Reais per US dollar, and since then it has declined to around 2 R$/US$.

Competitive Electricity Markets

554

Rationing 700,00 600,00

R$/MWh

500,00 400,00 Northeast drought 300,00 Drought in the south

SE/CW South NE North

200,00 100,00

05 20 /8/ 26

7/2

/20

05

04 20 /7/ 22

4/1

/20

04

03 20 /6/ 18

/20 /11 30

14

/5/

20

02

02

01 26

/10

/20

01 /20 9/4

21

/9/

20

00

–

Date

Fig. 15.4. MAE prices for heavy loads, 2000–2005. Source: Araújo (2006).

EBL

However, many more thermal plants would have to be built to change the context; and under present conditions, investing in thermal plants is a risky business. “Inﬂexible” thermal power, i.e., a thermal plant that could require to be dispatched, was created as an attempt to remedy this; but the inﬂexibility mechanism failed to consider that system price in normal conditions would not cover thermal operating (or even fuel) costs. It was also assumed that bilateral power purchase agreements would develop and create incentives for investing in thermal plants. This second scheme had two problems: one, it was rather difﬁcult to ﬁnd prospective buyers for an energy, the expected cost of which was higher than that of hydropower plants; second, securing fuel supply at reasonable conditions was less easy than expected, especially for natural gas (this remains an issue of concern). As a result, most thermal plants built until 2001 were owned by self-producers who already used gas as an industrial fuel and worried about power supply security, or else had Petrobrás as a partner. On the other hand, investment in hydropower plants was affected by a combination of factors: ﬁrst, the prospect of divestment for the existing hydropower plants did not encourage private investors to face the large risks involved in building new plants; second, divestment of existing plants was much more difﬁcult than expected, because of the wide range of stakeholder interests in reservoirs; and investment by federal generators was curtailed to cut public deﬁcit. The net result of hurdles for thermal and hydropower investment was the 2001 rationing crisis; more details may be found in Araújo (2001, 2006). The second reform, enacted in March 2004, had the basic goal to attract private investment in the expansion of electricity supply, in view of growth prospects for consumption and the perception that the State could not invest at the level required to attend demand. This was done through contracting and auctioning mechanisms, to be described below.

Reform of the Reforms in Brazil

555

Chief among issues faced by MAE was the design of the ﬁrst reform itself. The insufﬁciency of the Brazilian spot market to act as a signal for investment required the use of hedging mechanisms like forward contracts and power purchase agreements. As discussed above, the characteristics of the Brazilian power system – predominantly hydro-based, with large reservoirs that allow hydropower plants to be the marginal producer between 80% and 90% of the time – posed serious obstacles both to using ﬁnancial hedge instruments and to developing long-term bilateral contracts. As argued above, in such a system the volatility (understood as suddenness and size of price changes) of spot prices, coupled with extremely low prices during very long and uncertain periods (due to the size of reservoirs), effectively blocks the smooth operation of ﬁnancial hedges (especially given the immaturity of Brazilian ﬁnancial markets) and contracting. Without such hedges, for this system, reliance on the spot price as the main signaller for investment will lead to “stop and go” behavior (particularly because of the maturation lag required by hydropower plant investment); that is to say, in such a system spot prices will only be effective as an inducer to investment when the system is at risk of collapse. Another mechanism would be required for long-term contracts to be viable as an effective inducer to investment while keeping an adequate security margin for the system, until thermal participation grows enough for thermal plants to be the marginal producer for a signiﬁcant fraction of the time. The importance of the bulk market for a proper working of the power market, and the concern with the need for adequate stimuli to investment in generation expansion, led to the second electricity reform, enacted in March 2004 through Laws 10847 and 1084, with their attendant decrees14 . The next section will examine the institution in charge of the new trading arrangements.

EBL

15.4. The Second Reform: CCEE, Agents, and Contract Environments 15.4.1. CCEE and the regulated environment (ACR) At present, the Brazilian ESI is composed of a wide variety of agents, numbering more than 1250 (ANEEL, 2007): generators, large and small; thermal power plants burning fossil fuels, used to complement the system or supplying isolated subsystems; independent producers, mostly with gas-ﬁred power plants; transmission companies; large consumers; power traders; and distribution utilities that tend the wires business and serve captive consumers. Some existed before reform, others were created by the ﬁrst reform. Table 15.2 summarizes the ESI structure in 2005. In addition, the National Agency for Electric Energy (ANEEL) regulates and monitors the working of the ESI; the National System Operator (ONS) does the centralized physical dispatching; and the Trading Chamber for Electric Energy (CCEE) buys and sells electric power in the Regulated Contracting Environment (ACR), besides registering contracts made in the Free Contracting Environment (ACL). Above these bodies, there is also the inter-ministerial National Council for Energy Policy (CNPE) and the Monitoring Committee for the Power Sector (CMSE). The Ministry of Mines and Energy (MME) makes the plan and grants concessions. 14 The New Model for the Power Sector was enacted by laws 10.847 and 108.48 of 15 March 2006 and by decrees 5.163 of 30 July 2004; 5.175 of 9 August 2004; 5.177 of 12 August 2004; and 5.184 of 16 August 2004.

Competitive Electricity Markets

556

Table 15.2. Structure of the Brazilian ESI in 2005 Type

Activities

Enterprises

Binational

Generation

Itaipu

Federal ownership

Holding Generation Generation, Transmission Generation, Transmission, Distribution Distribution Nuclear Engineering Research

Eletrobrás Eletronuclear, CGTEE Furnas, Chesf, Eletrosul∗ Eletronorte Boa Vista, Manaus† NUCLEN CEPEL

State ownership

Generation, Transmission, Distribution Transmission Distribution

CESP, CEMIG, COPEL, CEEE Transmissão Paulista∗∗ 11 companies

Municipal

Distribution

5 companies

Generation Distribution

23 companies§ 40 companies†

‡

Private

Source: Araújo and Losekann (2001), updated by Araújo with ANEEL data (see Araújo, 2006). ∗ Between 1998 and 2004 Eletrosul was limited to transmission. † Five DISTCOS have been returned to government hands and are being temporarily managed by Eletrobrás: Ceal (AL), Ceam (AM), Cepisa (PI), Ceron (RO), and Eletroacre (AC). Eletronorte also controls Boa Vista Energia and Manaus Energia. ‡ There are also 423 free consumers, 63 independent power producers, 13 self-producers, 1 importer/exporter, and 44 energy traders registered with CCEE, of a total 608 market participants (CCEE data from 7 September 2005). § Including special purpose societies and joint ventures. **Sold in 2006 to the Colombian group ISA.

EBL

Chart 15.1 shows the relations between these organisms. Arrows denote hierarchy, and dotted lines connection with partial delegation of functions.

CNPE (Interministerial council for energy policy) CONGRESS MME (Ministry of mines and energy) CMSE (Monitoring committee for the power sector)

ANEEL (Electricity regulator)

ONS (Power system operator)

Chart 15.1. Federal organisms in the brazilian ESI.

EPE (Energy planning studies)

CCEE (Chamber for electricity trading)

Reform of the Reforms in Brazil

557

In this set-up, CCEE is charged with carrying out auctions for buying and selling electric energy, registering short- and long-term contracts, and with the accounting and settlement of short-term pool contracts. The present format of the pool results from a series of changes made to the former Bulk Electricity Market (MAE), replaced by CCEE after running into serious problems of design and operation noted above. From March 2004 onwards, electricity policy changed substantially, primarily to attract investment for sustained development of the sector. It had ﬁve explicit purposes: (i) to build a stable regulatory setup; (ii) to guarantee security of supply; (iii) to achieve fair tariffs; iv) to respect contracts; and (v) reintroduce planning in order to cope with demand growth. To this end, mechanisms were inserted into the market to enhance security of supply, among which: (a) a requirement that distribution companies contract for 100% of their forecast demand over a ﬁve-year horizon; (b) build realistic estimates for guaranteed energy of plants; (c) contracting hydropower and thermal plants in a mix that balances guarantee and cost; and (d) permanent monitoring of the security of supply, in order to have early detection of imbalances between supply and demand and to take steps to restore security of supply at least cost to consumers. To attract investment in generation, energy auctions for long-term energy contracts (15 and 30 years) were created to direct energy contracting by distribution utilities. This scheme aims at reducing risks for investors, while the auction by least price stimulates economic efﬁciency and in principle gives correct signals for the system expansion cost through competition. These auctions contemplate blocks of hydro and thermal plants, auctioned separately in order to obviate the issue of thermal investment in a hydro-based system. A peculiar characteristic of this arrangement is that all distribution companies form a pool that contracts with each auction winner; in effect, each contract is split among distribution companies according to their share of the market, one contract with the generator for each. This aims at spreading risks and beneﬁts as well as leveling supply tariffs. Finally, these auctions are fed by governmental planning studies to propose feasible (and with a preliminary environmental licence) expansion projects for a forecast demand growth, together with demand forecasts by distribution companies, although investors may propose alternative projects in the auction blocks. The auction system (see also Araújo, 2006) is aimed at the Regulated Contracting Environment (ACR), i.e., relating to the service of captive consumers by distribution utilities, and its aim is to ensure the provision of energy to these consumers in a reliable, equitable, and economically efﬁcient way (fair tariffs) through auctions and regulated pooled contracts. There also is a Free Contracting Environment (ACL), in which energy is contracted for free consumers through freely negotiated bilateral contracts. To deal with both environments, the Chamber of Electricity Trading (CCEE) was created. This institution replaces MAE, absorbing its functions and incorporating all its organizational and operational structures, with the following aims: (i) to administer the contracting of energy sales and purchases of public distribution utilities; (ii) to conduct energy purchase auctions for distribution utilities, under authorization of ANEEL; and (iii) to perform the functions of accounting and settlement in the two contracting environments of the market, ACR and ACL. CCEE is a private, not-for-proﬁt entity, with governance structure similar to that of MAE, having ﬁve councillors in its administration board; the President of the Board is nominated by MME and has vetoing power in deliberations that conﬂict with governmental policy or directives. Three councillors are nominated respectively by generators,

EBL

558

Competitive Electricity Markets

distribution utilities, and power traders, and one is elected by all agents. It is funded through contributions of its associates, which do not carry through to consumer tariffs. CCEE administers energy contracting under ACR and mediates the bilateral supply contracts each generator signs with each distribution utility. In this way it calculates the supply tariff to distribution utilities, to be used by ANEEL in deﬁning the captive consumer tariffs. CCEE also mediates the supply guarantee contracts each distribution utility has to sign, in order to reduce defaulting risks. Regarding the short-term (spot) market, CCEE takes on the same functions of MAE, i.e., accounting and settling differences between the amounts of energy contracted and those effectively consumed or produced by the agents, according to Energy Contracting Procedures homologated by ANEEL. In this market, every contractual difference is accounted for and ﬁnancial settlement is made monthly, being based on the Differences Settlement Price (PLD), which is published by CCEE in advance and was known as PMAE under the MAE. PLD is weekly calculated and published by CCEE, having for base the system marginal operational cost with lower and upper price bounds. The upper bound for PLD is deﬁned as the variable operation cost of the dearest thermal generation existing in the centralized dispatching program, and the lower bound is established by ANEEL considering hydropower plant operation and maintenance costs as well as ﬁnancial compensation for the use of water resources. In this way, for the purposes of accounting and settlement, contractual differences must be valued at PLD and settled monthly. Nevertheless, gains, losses, and penalties from contractual deviations of distribution utilities are the object of a yearly conciliation, taking into account the effects of seasonal variation in consumption as well as atypical intra-annual variations that might require compensation. Distribution utilities must prove that their measured market is 100% covered by contracts. When a utility is overcontracted, difference settlement will produce revenue gains or losses, in case the monthly PLD is higher or lower than the contracted purchase price in ACR. Allocation of these gains and losses must obey the following rules:

EBL

•

Over contracting up to 3% of the market: gains will be appropriated by the utility and losses will be passed through to consumers next year. • Over contracting above 3% of the market: gains and losses are absorbed by the utility. As mentioned earlier, in regulated long-term energy trading contracts (CCEAR), the auction winners of ACR energy auctions must sign bilateral contracts with all distribution utilities, proportionally to their stated requirements. The sole exception is the “Adjustment Auctions” for up to one year ahead, where contracts are signed between one seller and one distribution utility only and may last no more than two years. Auctions aimed at new generation exist for contracts with delivery ﬁve (“A-5”) or three (“A-3”) years ahead, with durations between 15 (thermal) and 30 (hydropower plants) years; contracts from “A-1” auctions, aimed at existing plants, have durations from 5 to 15 years. There are two modalities of regulated long-term contracts: by amount of energy to be delivered, in which the seller takes all hydrologic risks, and by energy availability, in which the buyers take on hydrologic risks (and can buy energy at a cheaper price). For regulated long-term contracts involving energy from existing plants, the rules deﬁne three possibilities for decreasing amounts contracted: •

Compensation for the exit of free consumers – distribution utilities, after compensation for surplus and deﬁcit (part of the Trading Convention), may reduce their contracts for the uncompensated balance due to the exit of free consumers.

Reform of the Reforms in Brazil

559

•

Reduction, at the demand of a distribution utility, of up to 4% of the amount contracted, to adapt to market deviations of demand forecasts. • Reductions due to increases in bilateral contracts signed up to 16 March 2004. These concern contracts that started with small amounts, but with clauses increasing energy transacted, mainly as new plant entered operation. It should be noted that reductions are applied uniformly to all regulated LT contracts of the distribution utility with existing plants derived from ACR auctions. Note also that distribution utilities may pass through contracted amounts up to 103% of their load [in energy; we stress once again, as in Araújo (2006), that in a hydro-based system with large reservoirs energy rather than capacity is the limiting factor]; this acknowledges that a perfect forecast does not exist and allows a tolerance for forecast errors. Finally, when contracts from “A-3” auctions exceed 2% of demand, pass-through is limited to the least cost contracts from “A-5” and “A-3” auctions. These measures are justiﬁed to avoid inefﬁcient outcomes. In effect, if a distribution utility contracts an excessive amount of energy from new plants, and later reduces energy from existing plants, this would be a socially inefﬁcient use of resources. Therefore, in order to provide correct stimuli, if the purchase of energy from existing plant is less than the lower bound for contracts the pass-through of the cost of energy purchased from new plants will have an upper bound. This is done as follows: distribution utilities have to contract 100% of their market through auctions; if after a 12-month period the utility ﬁnds itself short of energy, it will have to buy energy through short-term contracts, with prices subject to the vagaries of the spot market. These prices may not be passed through to ﬁnal consumers above the limit of 103% of the tariff. As mentioned above and as discussed by Correia et al. (2005), energy supply contracts in ACR may vary only between two modalities15 :

EBL

•

Contracts for delivered energy, similar to initial contracts, in which all risks are taken by generators to supply the energy contracted. • Contracts for energy availability, in which all risks of production deviations relative to assured energy are assigned to the pool and passed through to captive consumers. To the general rule that contracts in ACR are formalized in bilateral contracts between each generator and all distribution companies, there are three exceptions: •

The Binational Itaipu plant, the energy of which is traded by Eletrobrás, only for consumers in the South, Southeast, and Center-West. • The Incentives Program for Alternative Energy Sources (PROINFA)16 , the energy of which is also traded by Eletrobrás for distribution companies in the whole National Interconnected System. • Distributed generation17 ; in order to alleviate the requirement for transmission investments, distribution companies are allowed to contract energy from distributed 15

Note that all contracts in the distribution pool refer only to distribution companies in the National Interconnected System, which represent 98% of the market. 16 PROINFA was created to diversify the supply of electricity, expanding the generation share of small hydropower plants, thermal plants burning biomass, and wind parks in its ﬁrst phase, although other sources may be contemplated in the future. The ﬁrst phase is to have 144 power generation projects, totalling 3300 MW of installed capacity. 17 Article 14 of decree 5163/2004 qualiﬁes as distributed generation in Brazil the production of electric power from generation plants directed connected to the buyer’s distribution system, except that from

560

Competitive Electricity Markets generation plants. This must be done through public bidding directly promoted by the distribution utility, and may not exceed 10% of its load.

The contract pool format that characterizes ACR leads to a single reference tariff for the pool at each auction. That is, all distribution companies participating in an auction purchase energy at the average price of the total amount of energy contracted in the auction. However, tariffs applied by distribution utilities to their consumers are different in function of the composition of their contract portfolios. These may be composed by contracts signed in distinct auctions with distinct average prices, by initial contracts until January 2006, by contracts with energy from Itaipu, from PROINFA plants, and from distributed generation. Other cost factors also have an inﬂuence in differentiating energy tariffs among distribution utilities, such as the tariff review methodology employed by ANEEL, which considers socio-economic characteristics of the market to determine costs and investments to be remunerated by the tariff 18 , the market share of ﬁnal consumers in the Low Income Household category19 and whether state taxes exist that impinge on tariffs according to the geographical situation of the utility. 15.4.2. The free (non-regulated) contracting environment for bilateral contracts (ACL) Another form of contracting in the Brazilian power system is through free bilateral contracts. Given the characteristics of the Brazilian power system, as discussed in Section 15.1, contracts are a useful tool to reduce agent exposition to price uncertainty. Both Brazilian electricity reforms have paid attention to bilateral contracting. In the ﬁrst reform, ANEEL assigned “initial contracts” in December 1998. These were middle- to long-term bilateral contracts signed between generators and distribution utilities, based on “assured” generation and programmed load demands until December 2005, with prices deﬁned by ANEEL, and aiming to provide a smooth transition to free contracting. To this end, they were staggered: starting in January 2003, amounts contracted would be reduced by 25% yearly, and by January 2006 there would be no more initial contracts. As Pires (2000) reminds us:

EBL

Immediate passage to free energy supply contracting in MAE would cause an electricity price shock in Brazil, since the average supply cost, to be passed through to (I) hydropower plants with nameplate capacity above 30 MW; or (II) thermal power plants, including cogeneration, with energy efﬁciency below 75%. 18 ANEEL divides utility costs in two parts: part A (uncontrollable costs) and part B (controllable costs). Part A costs are always passed through to consumers and comprise energy purchase costs, sector charges, taxes, and energy transport. Part B comprises operational costs, investment remuneration, and depreciation. To estimate O&M costs, ANEEL constructs a reference enterprise for each utility. This reference enterprise is used as a benchmark for cost efﬁciency; in other words, O&M costs considered in the deﬁnition of distribution tariffs are not measured utility costs, but rather those considered to be adequate for serving its market, taking into account socio-economic features of that market. In the case of investment and depreciation costs, ANEEL estimates a remuneration base for each utility, in which it includes assets and investment judged to be prudent and necessary for efﬁcient service. 19 These are taken to be consumers served by one-phase circuits, consuming less than 80 kWh/month over 12 months, without consuming above 120 kWh/month for two or more months (to exclude summer residences) or consuming between 80 and 220 kWh/month and registered as low-income households to receive a family stipend from social programs (ANEEL resolutions 246/2002, 485/2002, 694/2003, 044/2004).

Reform of the Reforms in Brazil

561

distribution utilities and from these to ﬁnal consumers, would reﬂect the growing system marginal expansion costs. Aiming to mitigate the effects of a possible tariff shock in the beginning of the operation of the new model20 , the government established a transition period in which initial contracts are being constituted between generators and buyers; in practice, these do nothing more than reproduce the same clauses present in existing supply contracts. Thus, only the amount of energy not involved in initial contracts could be freely traded, either in the MAE spot market or through bilateral contracts. In order to provide a learning period, law 9648 of 27 May 1998 established that an initial phase, in which competition among generators would be restricted to energy from new projects, would be followed by a transition period, from January 2003 to January 2006, where initial contracts would gradually free energy by 25% a year. By 2006, full competition in generation was expected. For distribution utilities, liberation from initial contracts meant the possibility to contract up to 15% of their market in the bulk market, and sign new bilateral contracts for the remaining 85% by legal requirement. Later on, through the ANEEL resolution 511 of 12 December 2002, this requirement was increased to 95%21 . The framework did not establish that agents had to sign long-term contracts between them: the same resolution (cf. footnote 21) only considers contracts with at least two years duration. The development of longer-term contracts was expected to arise from expectations regarding the evolution of electricity prices. Legislation also guaranteed the pass-through of energy costs to ﬁnal consumer tariffs up to a certain value, the Normative Value (VN) that varied according to plant technology. The logic behind this was an attempt to induce efﬁciency through a price cap on contract prices, limiting the pass-through of energy purchasing costs to captive consumers. Originally, VN were estimates by ANEEL of long-term marginal generation costs according to the generation technology; but in 1999, pass-through was extended to VN + 11.5% in view of many complaints. The assumption was that competition among generators would force prices down to competitive levels. Nevertheless, in practice the VN + 11.5% value became a ﬁxed price level rather than a price ceiling, due to the low level of competition in generation, contrary to expectations implicit in the Cardoso reform framework. Lack of competition in the generation segment was due to several factors; notably, the malfunction of MAE as price signaller and the inconclusive privatization of the model. The ﬁrst reform initially foresaw the splitting of large state-owned generators in order to increase the number of generators, and later to privatize them. This would be complemented by the unbundling of activities, to allow competition in potentially competitive activities and put an end to cross-subsidies, prior to divestment. Privatization started with the two distribution companies absorbed by the Eletrobrás Group – Light and Escelsa, sold

EBL

20

This refers to the framework of the ﬁrst electricity reform, led by the Cardoso administration. In its Article 6, this resolution stated: “Of the amount of energy traded by agents participating in MAE with ﬁnal consumers, at least 85% (eighty-ﬁve per cent) must be guaranteed by own plants or by power purchase contracts with duration 2 (two) years or more in any submarket, and at least 10% (ten per cent) guaranteed by assured energy of own plants or by bilateral contracts of any duration in any submarket, totalling 95% (ninety-ﬁve per cent) of the amount of energy traded.” The same bound was applicable to free consumers, according to Section 1 of the same article: “Free consumers which are MAE members must prove that at least 95% of the energy consumed is guaranteed by assured energy or effective generation of own plants or bilateral contracts of any duration.” 21

562

Competitive Electricity Markets

off in 1995 and 1996, respectively, before sector regulation had been set up22 . Next, the government intended to sell the four big generators controlled by Eletrobrás (Furnas, Chesf, Eletronorte, and Eletrosul), but these proved harder to privatize for the reasons discussed in Sections 15.1 and 15.2; instead, state-level distribution companies were divested for two reasons: they had a captive market, which made them easy to sell, and their divestment would improve the value of generators, since those distribution companies were in debt to the large generators. As Leal (1998, p. 9) argues, “since state-level distribution companies in general showed among their liabilities a high degree of indebtedness towards energy suppliers, it made sense to privatize them beforehand, so that these bad credits, present and future, would not substantially reduce the economic value of generators.” In the end, among federal generators, only the generation assets of Eletrosul were privatized, being sold off to Tractebel (of the Suez Group). Therefore, given the feeble competition in generation for contracts with distribution utilities, VN (or, rather, VN + 11.5% as argued above) became the price around which a good part of bilateral contracts were signed in the period; this became quite visible in selfdealing contracts. The Cardoso reform framework set a limit for this kind of contracts, since the pre-reform industry structure was very concentrated and vertically integrated. Before the reforms of the 1990s, distribution was carried out by 55 companies, of which 3 were federal (Light, Escelsa, and Eletronorte, which was integrated), 23 owned by Federation states, 5 municipal, and 24 private (these latter served mostly small towns, representing only 5% of the market)23 . Among state-owned distribution companies, 5 were notable for being vertically integrated and holding among them 32% of the installed capacity in the Brazilian generation park: CEEE (Rio Grande do Sul), CEMIG (Minas Gerais), COPEL (Paraná), CESP, and Eletropaulo (São Paulo) (Pinto Jr., 1993). This was due to the fact that these companies, plus Light for Rio de Janeiro, served the biggest urban centres in the country and needed own generation to keep acceptable reliability levels in the system. Some but not all companies were restructured before being privatized, and some were never divested. Only CESP, Eletropaulo, and CEEE were restructured, and their distribution assets were privatized24 . In view of this situation, an additional mechanism was sought to limit the possibility for distribution utilities holding generation assets to contract their own energy at strategic prices. An upper bound to self-dealing would also signal against vertical reintegration through power purchase agreements, while allowing an agent to invest both in distribution and in generation. Nevertheless, inconsistent signals were emitted by the government on self-dealing within the economic groups taking shape in Brazil, with the entry of important foreign groups via privatization of power distribution utilities. In some cases, government measures even induced an increase, rather than decrease, of generation investment by distribution utilities. First of all, integration of activities was seen as a means to raise the market value of distribution companies to be privatized. Therefore, some divestment auctions, like those

EBL

22

This inverted the “textbook” model, despite a general consensus that it would have been best to do otherwise. Reasons for the inversion were mainly the attractiveness to buyers of distribution companies with a captive market, and the desire to show results of the restructuring program. 23 Rodriguez-Pardina and Estache (1996), updated by Araújo and Oliveira (1997). 24 CESP was split into a state-owned transmission company (CTEEP, which was only privatized in July 2006) and three generators, two of which were privatized and one could not be sold, retaining the name CESP.

Reform of the Reforms in Brazil

563

of CELPE, COSERN, and COELBA, had clauses to stimulate investment in generation and transmission in order to raise the value of the assets to be privatized. On the other hand, ANEEL tried to establish limits to market concentration through resolutions 94/1998 and 278/2000, which deﬁned concentration limits in order to restrain mergers and acquisitions and leave a reasonable number of agents. As Pires and Piccinini (1999) note, resolution 94/98 forbade market agents to: (a) hold more than 20% of the national installed capacity or 25% and 35%, respectively, of the existing capacity in the interconnected subsystems South/Southeast/Center-West and North/Northeast; (b) hold more than 20% of the national distribution market or 25% and 35%, respectively, of the distribution market in the interconnected subsystems South/Southeast/Center-West and North/Northeast; and (c) hold cross-shares in generation and distribution leading to a market share above 30%, calculated as the arithmetic sum of the shares in each market. Thus, a distribution company can only buy energy produced by itself up to 30% of its demand. In view of the latent threat of energy rationing due to low investments in generation and the possibility of an unfavorable rainy season, ANEEL resolution 278/00, in its Article 7, excluded thermal plants in the Priority Program for Thermal Plants (PPT) 25 from the upper 30% bound. In this case, investors in these thermal plants could have unlimited participation in the distribution market. This explains why some groups invested in thermal power plants through their distribution companies. Thus, the freeing of initial contracts after January 2003 opened way to self-dealing using the bounds of resolution 278/00, since Law 10604/2002 forced contracting between generators and distribution utilities after that date through auctions, but allowed selfdealing in the former terms. For these self-dealing contracts, price was as a rule set by the ceiling VN (or, rather, VN + 11.5%, as argued before). This price was even more advantageous for distribution companies producing electricity through thermal plants. Under the shock of the rationing crisis, in order to stimulate the investment in thermal power plants – both because they would increase supply security and have a shorter maturation lag than hydropower plants – ANEEL created a correction formula for VN with three factors, according to resolution 22/2001, article 90: K1 (correction by IGPM26 ), K2 (fuel price variation), and K3 (exchange rate variation). Since both K1 and K2 were affected by K3, readjustments were systematically higher than inﬂation indices in each period. Moreover, since these distribution companies saw in this a very attractive guarantee of revenues, they signed extremely long-term self-dealing contracts; thus, CELPE signed a 20-year supply contract with its thermal power plant, beginning December 2003. Since the cost of thermal plants is much higher than the marginal system cost 80–90% of the time, the energy price that consumers had to pay with these schemes was signiﬁcantly higher than the average system cost. The problem was bigger where self-contracting answered for a larger share of a utility’s market, and was aggravated by the PPT. Considering that the rules established during the implementation of the Cardoso reform had stimulated self-contracting at prices detrimental to consumers, in view of the methodology to calculate VN and its readjustments, the Lula reform forbade self-dealing or bilateral contracting within a single economic group (articles 20 and 30, law 10848/2004)

EBL

25

As already mentioned, this was a program launched by the Federal Government in February 2000 with the aim of avoiding hydro plant reservoir depletion in the face of high demand growth in relation to investment. 26 This was a bulk price index that had been used extensively in power regulatory contracts.

564

Competitive Electricity Markets

and extinguished VN and its readjustments based on K1, K2, and K3 (paragraph 40 of article 20, law 10848/2004). To protect captive consumers from the results of the earlier experiment, the Lula reform deﬁned that the purchase criterion should be minimal tariff or price (item 70 of article 20, decree 5163/2004) and forced all purchase of energy traded by distribution companies to be done through public and transparent auctions (article 20, law 10848). To this end, the new model split the Brazilian market into two trading environments; captive consumers would be sheltered in the regulated environment (ACR), with competition for bilateral contracts and effective freedom of large consumers in the free environment (ACL). Since the reform was not aimed at making a clean break with its predecessor but to consolidate institutions, all contracts signed till then received guarantees that they would be respected, including those of self-dealing; all of them will be administered within ACL until their expiration. Agents are free to make bilateral contracts in ACL, deﬁning prices, quantities, durations, and hedge clauses, with one important exception: State-owned generators, even when contracting in ACL, must do it through public auctions approved and supervised by ANEEL. Besides free consumers, ACL may comprise concession holding generators, independent producers, power traders, and power importers. This market represents nowadays over one-fourth of all transactions and tends to focus on short-term contracts. This reﬂects the relative energy glut of the moment, but may change if supply becomes tighter. 15.5. Electric Power Competitive Auctions in Brazil

EBL

According to the present Brazilian regulatory framework established by Law 10848 of 15 March 2004, distribution utilities must ensure that their market demand has full contract coverage by purchasing electric energy through public auctions conducted within ACR. Brieﬂy stated, auctions may be deﬁned as a bargaining mechanism having as chief characteristic to lead quickly to the revelation of the price of a given good, though its value may be unknown. To this end, they must set limits to room for strategic actions by agents and stimulate the revelation of opportunity costs and of expectations for the future behavior of supply and demand. Additionally, the efﬁciency of an auction will depend on the existence of speciﬁc mechanisms and rules that increase its attractiveness and reduce possibilities for collusion, predatory competition, and other forms of market power (Klemperer, 2004). Auctions may then be understood as a space for regulated competition, with rules and institutions that make competition more transparent and minimize the use of market power. The existence of an ofﬁcial market for electric energy, operating through public auctions, may thus operate as an important instrument to consolidate the liberalization process of the ESI in Brazil. This role of reference market could be played by transactions in the spot market made through daily auctions; nevertheless, speciﬁcities of electricity and of the Brazilian market hinder this approach, while, as shown in Adib et al. in Chapter 7 of this volume, antitrust law scarcely addresses the market power problems encountered in electricity markets. Electricity has speciﬁc traits that distinguish it from other goods (Lee, 2004); in the ﬁrst place, it is a critical resource for various activities regarding the well-being of the population, there being thus a critical consumption threshold below which serious social disruptions may occur. It is also a local good having signiﬁcant restrictions for long distance transport, and it is not easily storable, so that production and consumption must be simultaneous. Thus, the correct price for electric energy may only be known ex post, which reduces the efﬁciency of demand adjustment in function of price elasticity. Finally,

Reform of the Reforms in Brazil

565

it shows large hourly and seasonal consumption variations; this implies that physical networks are dimensioned according to peak demand and that part of generation capacity remains idle most of the time. In consequence, the supply curve of electric energy is inherently very steep when approaching the supply capacity and some volatility with price spikes cannot be avoided. Besides, new investment normally means maturation lags above three years; in countries like Brazil, with a high potential demand growth, this implies a relatively high risk of recurring supply scarcity, with rationing and blackouts. This systemic risk is even greater in hydropower-based systems, since these have an additional uncertainty for rainfall and thus for energy available in reservoirs. Finally, another relevant feature of the Brazilian supply industry is the dominant role of public agents in the generation activity. After privatizing distribution assets, and after the 2001 rationing, the Brazilian government effectively paralyzed the federal generation divestment process (tacitly under the Cardoso administration, and explicitly in the Lula reform), with a view to attract existing capital to the expansion of the system. The new institutional framework was thus designed to allow public and private ﬁrms to coexist in a competitive environment. Auctions play a central role in this design, since the existence of clear rules and a transparent trading process work as a guarantee to private ﬁrms against a possible abuse of power by public ﬁrms. Auctions for generation projects were thus instituted in Brazil with the following purposes: •

Create a long-term contract market that generates efﬁcient and timely price signals to guide the expansion of installed capacity. • Ensure that the purchase of energy to supply captive consumers is done in a competitive and transparent way, leading to fair tariffs. • Provide clear, easy to audit trading rules, in order to hinder collusion and the use of market power to manipulate prices.

EBL

In this context, ACR may be considered as a buyers’ pool that aggregates the demand of the various distribution utilities in periodic auctions, in order for utilities to sign bilateral contracts with the generators who offer winning bids in each auction. By pooling utility demands into a monopsony, ACR is expected to achieve gains of scale; since each generator signs contracts with every utility, risks are spread out among sellers (as well as among buyers). Electricity supply auctions and contracts may take three forms: contracts for energy from existing plants, contracts for energy from new plants, and adjustment contracts. The process and its results are discussed at length in Araújo et al. (2007). 15.5.1. Auctions to purchase electricity from existing plants At ﬁrst, auctions of “existing energy” were formatted to allow contracting surplus installed capacity (due to progressive ending of initial contracts) before auctioning energy from plants yet to be built. In order to improve decision making by sellers and provide exchange of information on the value of electricity, these auctions had a ﬁrst phase with open bidding, where agents modiﬁed the quantity offered according to current price announced, and a ﬁnal phase with sealed bids and discriminatory prices. To protect against collusion, these auctions had a secret reserve price and some demand might thus remain without contracts. Taken as a whole, auctions to purchase energy from existing plants accomplished their purpose since they allowed early contracting of a large volume of surplus electric energy,

Competitive Electricity Markets

566

restoring an acceptable level of risks and uncertainties and thus creating room for new investments. Besides, trading rules showed that public and private ﬁrms could coexist in a competitive and transparent market. The results of these auctions, concerning prices and amount of energy contracted, may be followed in Fig. 15.5 15.5.2. Auctions to purchase energy from new plants Considering the Brazilian experience, the “missing money” or resource adequacy problem, referred in Part 3 of this book, is addressed through long-term contracts resulting from speciﬁc auctions for greenﬁeld power plants. These auctions are carried out ﬁve and three years before the year where energy is needed, and are known as A-5 and A-3 auctions. The former ones regard generating plants that can start operation in ﬁve years time, mostly hydropower plants; the latter ones, plants that can operate within three years, mostly thermal plants. This is intended to allow contracting a portfolio of plants that efﬁciently combine ﬁxed capital costs (higher in hydro and nuclear plants) and variable costs (higher in conventional thermal plants), and permit an optimal dispatching according to hydrological context. Contracts signed to purchase new generation must have duration between 15 and 35 years and have clauses to stimulate efﬁcient pricing. Two contract modalities exist, upon decision by MME: Contracts for energy volumes, where sellers take on all the risks, and contracts for energy availability, where the pool takes on the risks and passes costs and beneﬁts through to ﬁnal consumers (see also Section 15.4.1). Energy availability contracts signed in Brazil foresee capacity payment clauses but may not be understood as capacity market mechanisms, as they are restricted to thermoelectric plants and aim at dealing with fuel price volatility and dispatch uncertainties. Therefore, they constitute rather an instrument to confer predictability to the cash ﬂows of thermoelectric plants that have signed long-term contracts.

EBL

10.000

120.00 9.054

9.000 97.25 8.000

80.00 72.26

6.000 61.72

60.00

64.50

4.000 40.00

3.000 2.000 1.172

1.325

1.166

2007 (1st Auction)

2008 (2nd Auction)

2009 (4th Auction)

20.00

1.000 102 0.00

–

Supply in

2005 (1st Auction)

2006 (1st Auction)

2006 (3rd Auction)

Energy purchased

Price

Fig. 15.5. Auctions to contract energy from existing plants. Source: CCEE.

R$/MWh

Average MW

80.98

6.782

7.000

5.000

100.00

86.89

Reform of the Reforms in Brazil

567

Hydropower plants, on the other hand, have only contracts for energy amounts and must incorporate risk premiums in their selling prices in order to cope with additional reliability costs. This approach recognizes that generators possess better pricing mechanisms for exposure risks and that competition in long-term auctions enables the valuation of the utility of reserve capacity necessary to ensure supply of a speciﬁc amount of energy. In addition to that, in order to match thermal availability contracts and hydrocontracts for energy amounts, ANEEL employs mathematical programming models to calculate an expected dispatch rate and the amount of energy effectively aggregated to the market. Nevertheless, auctions designed to buy energy from new plants enabled expansion of generation capacity through market signals in a time horizon compatible to maturation lags required by new investments, which are important to mitigate the risk of long periods of scarcity. Figure 15.6 sums up the results of the ﬁrst four auctions that took place in Brazil. 15.5.3. Adjustment auctions In accordance with ACR trading rules, uncertainty as to market growth is absorbed by distribution companies, which must declare their expected demand and contract the necessary energy in auctions 1 to 5 years before the date. In order to allow proper management of this risk, there exist two mechanisms for complementary contracts: Adjustment Auctions for Electric Energy Purchase and the market for Settlement of Differences. Adjustment Contracts are utility-speciﬁc and made through public auctions authorized by ANEEL, one to two years before the date the energy will be needed, and last no more than two years. The pass-through of prices of these contracts has an upper bound in the Reference Value (VR) for the current year. In the present framework, VR is the price resulting from pool contracts in A-5 and A-3 auctions and due to start delivery in the year where the adjustment energy is required. There is thus a regulatory incentive for distribution companies to contract early on their expected market instead of risking a high exposure to adjustment auctions. Auctions for

EBL

140.00

1.028 1.000

134.29

134.25

889

855 131.25

800

130.00 126.77 644

600

123.68

561

125.00 120.00

117.78

116.04

115.00

400

200

110.00

108.59 71

105.00 46

–

100.00

Supply in 2008 (A-5) Hydro (Mwa)

2009 (A-5) Thermal (Mwa)

2009 (A-3)

2010 (A-5)

Hydro Price

Fig. 15.6. Auctions to purchase energy from new plants. Source: CCEE.

Thermal Price

R$/MWh

Average MW

135.00 862

Competitive Electricity Markets

568

new plants operate then as catalyzers to re-engage in energy planning in order to minimize investor uncertainty regarding future cash ﬂows. They also create conditions for a market in long-term contracts, which may be taken as guarantees of receivables to obtain more favorable interest rates. The ﬁrst Adjustment Auction was to have taken place on 31 August 2005, but was called off for lack of demand from the utilities. The sole Adjustment Auction carried out to date happened on 1 June 2006, with the participation of three distribution utilities from the Northern and Northeastern submarkets, CELB, CELPA, and SAELPA. Following the format used, the buyers informed the auctioneer of required energy in average MW, duration of contracts, and maximum acceptable price. Would-be sellers were invited to make quantity bids at the going price, which would vary until supply equated demand, characterizing a clock-auction design. The result of the Adjustment Auction is shown in Table 15.3.

15.5.4. Preliminary assessment Well-designed auctions can lead to the revelation of prices and costs of goods having uncertain value. Their format may vary in order to obtain better conditions for competition. Auctions should be as attractive as possible, and opportunities for collusion should be minimized. Of course, real conditions have to be considered in any performance assessment. One of the aims of the 2004 reform was to use surplus supply after the 2001 rationing to reduce the rate of growth of ﬁnal consumer tariffs, which had grown signiﬁcantly above inﬂation27 , especially for captive consumers, who had borne most of the costs and inconvenience of the 2001 rationing and the following sector crisis. To this end, the reform created the concept of “old” or “existing” energy, based on cheaper generation technologies and having a large part of costs already depreciated. This was a power play by the federal government to reduce electricity prices. Generators were not interested in this distinction, since they hoped to raise cash for future investments. A bargaining process began, the generators sending signals that auction prices should be high to ensure the attractiveness of power business in Brazil. From this point of view, the auctions carried out were reasonably successful because they broke the inﬂationary bias in prices and devolved to consumers’ part of the revenue extracted from them during the 2001 crisis. Besides, they signalled expected future prices

EBL

Table 15.3. Second adjustment auction for electric energy purchase Duration (months)

Average Price R$/MWH

Sub-Market Northeast

3 6

25 15

2912 3431

Sub-Market North

6

135

4563

Source: CCEE. 27

Quantity (Average MW)

Cf. Fig. 16.11 in Araújo (2006).

Reform of the Reforms in Brazil

569

500 450 400 1st New energy auction

350 300 250 200

2nd Existing energy auction

150 100 50

06 20 1/ 6 /1 00 08 9/2 6 /0 00 08 7/2 6 /0 00 08 5/2 6 /0 00 08 3/2 6 /0 00 08 1/2 5 /0 00 08 1/2 5 /1 00 08 9/2 5 /0 00 08 7/2 5 /0 00 08 5/2 5 /0 00 08 3/2 5 /0 00 08 1/2 4 /0 00 08 1/2 4 /1 00 08 9/2 4 /0 00 08 7/2 4 /0 00 08 5/2 4 /0 00 08 3/2 4 /0 00 08 1/2 3 /0 00 08 1/2 3 /1 00 08 9/2 3 /0 00 08 7/2 3 /0 00 08 5/2 3 /0 00 08 3/2 3 /0 00 08 1/2 /0

08

IBOVESPA

IEE

Fig. 15.7. Comparison between IBOVESPA and IEE. Source: Reuters.

for new energy (whether from “botox”28 plants or from expansion projects) compatible with investment costs and capital remuneration. This is shown by inspection of prices of energy to be delivered in 2009 in auctions for new plants, compared with prices for existing plant. Another independent check may be obtained through comparison of the behavior of stock market prices: After energy auctions, the IEE index (which aggregates all open capital electricity companies) cut loose from the aggregate stock market index IBOVESPA, growing 50% points above IBOVESPA since January 2003 (Fig. 15.7).

EBL

15.6. Conclusions A problem with the ﬁrst market-oriented reform of the Brazilian Electricity Supply Industry was the lack of adequate investment in expansion. One major cause for this, which led to the 2001 electricity crisis, was that trading arrangements exposed thermal plants to excessive risks. Another cause was that the organism created to deal with electricity trading became paralyzed through conﬂicts caused by incomplete rules and ﬂawed governance. As a consequence, the market did not produce strong and clear signals for potential investors. Another problem was that electricity prices increased faster than inﬂation, especially after the 1999 devaluation and the 2001 crisis. This made reform highly unpopular, and had several causes (cf. Araújo, 2006); one of these was that safeguards to prevent abuse of market power, and especially self-dealing at unfair terms for consumers, were crippled by ad hoc measures taken in a hurry to mitigate the 2001 power crisis. As discussed in Araújo (2006), the 2004 reform had several signiﬁcant traits; the bestknown of these are the greater role played by the central government and a relative centralization of trading arrangements. The new arrangements were designed to deal 28

This was a nickname given to plants built after the allocation of initial contracts, i.e., after 2000. Legally, they could compete in auctions aimed at new plants, hence the (ironic) epithet.

570

Competitive Electricity Markets

with the two problems mentioned above: to attract new investment in generation, both hydro and thermal, and to obtain fair prices for consumers with competition in expansion. Central to these arrangements were new contracting procedures between generators and distribution utilities, based upon auctions. This Chapter endeavored to discuss the rationale for them, and to analyze the performance of auctions. As we have seen, they appear to be moderately successful. Some ﬁnal considerations are, however, in order. Regarding auction design, special attention should be given to the use of available information. Full disclosure of information (like identifying agent bids) during an auction may facilitate collusion. To reduce prospects for cartel formation, information disclosure may thus be limited. This would also simplify decision making, since it would hinder strategies that are not exclusively based upon price signals and on individual preferences (Correia et al., 2006). Another observation is that electricity trading auctions appear to be accepted by the market, and are probably here to stay. It is also clear that rules and formats are still evolving, aiming at correcting ﬂaws and keeping uncertainty at a level that avoids strategic plays, especially in sequential auctions where market power and collusion would ﬁnd a favorable environment. Otherwise, there could be co-operative behavior in determining equilibrium price since there will be opportunities for learning and information exchanges, which may lead to investor co-ordination (Correia, 2004). Contract arrangements based on auctions appear to be a positive step toward solving the issue posed by Joskow in his question cited in the introduction to this chapter. This is however but one instrument among many. Several issues must still be solved to guarantee security of electricity supply. The single most urgent issue is the deﬁnition of efﬁcient, integrated procedures of municipal, state, and federal organisms to carry out the process of environmental licensing for new projects. Constant delays in granting operating licenses for hydropower plants or for gas pipelines, while granting easier terms for oil-burning power plants, may lead to more expensive and more polluting electricity. This would be sadly ironic, in view of the prospects for climate change (Cf. Chapter 14 by Ford in this volume) and the Brazilian potential for renewable power, especially hydropower. Financing restrictions remain, for the time being, another bottleneck for sustained supply expansion. The government expects that PPA’s with duration compatible with the concession period, and new ﬁnancing tools in the capital market, may leverage the resources required. Nevertheless, without a speciﬁc policy by the National Economic and Social Development Bank (BNDES) and the participation of private institutions through Project Financing, for instance, the ﬂow of resources may not be sufﬁcient. Aiming to remedy this, a special line of funding has been created for investment in electricity generation and transmission. In short, despite its merits, the present reform model still needs adjustments in many dimensions, some of which were covered in Araújo (2006). Also, as Chao et al. (Chapter 1 of this book) remind us, long-term contracts have their own built-in risks. Systemic risk may not be completely eliminated, and means to cope with it must be sought. There remain conﬂicts of competence between organs of direct and indirect public administration as well as between different regulatory agencies. There also are signiﬁcant gaps in the legislation dealing with natural gas and with isolated systems, which have strong impact on the power sector and may jeopardize the expansion of thermal generation. It is to be hoped that the Natural Gas Bill under discussion in Congress may soon be voted to become a law. Last but not least, special attention should be given to demand expectations signalled by the fast growth of the free contracting environment, which now represents 26% of the market. Up to now the government has mostly been concerned with the regulated

EBL

Reform of the Reforms in Brazil

571

environment, but future expansion will have to provide energy for the whole market, both regulated and free. As the country ﬁnancial markets become more mature, ﬁnancial hedge may become useful. The present setup may require further adjustments as circumstances change. References Adib, P., and Hurlbut, D. (this volume). Market power & market monitoring. Chapter 7. Adib, P., Schubert, E., and Oren, S. (this volume). Resource adequacy: Alternate perspectives and divergent paths. Chapter 9. Amundsen, E.S., Bergman, L., and von der Fehr, N.-H.M. (2006). The Nordic electricity market: robust by design? In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Elsevier. ANEEL (2007), Agentes no Mercado, http://www.aneel.gov.br/area.cfm?idArea=10 Araújo, J.L.R.H. (2001). Investment in the Brazilian ESI: What went wrong? What should be done? Workshop on Competition and Regulation: The Energy sector in Brazil and UK/EU, Oxford, St. Anne’s College, 4–5 June. Araújo, J.L.R.H. (2006). The case of Brazil: Reform by trial and error? In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Elsevier. Araújo, J.L.R.H. and Besnosik, R. (1992). Regulation, institutional structure and the performance of the Brazilian electricity sector. COPED Report, Rio de Janeiro. ´ Araújo, J.L.R.H and Losekann, L. (2001). Atualização da industria eléctrica Brasilerira. Research Report, Instituto de Economia da Universidade Federal do Rio de Janeiro. Araújo, J.L.R.H., Costa, A.M.A., Correia, T., and Melo, E. (2007). Energy contracting in Brazil and electricity prices. IAEE International Conference, Wellington, New Zealand, 18–21 February. Araújo, J.L.R.H. and Oliveira, A. (1997). The Brazilian electricity reform: Issues and perils. 18th Annual North American Conference of the USAEE/IAEE – International Energy Markets, Competition and Policy. San Francisco, California, 1383–9. Benth, F.E. and Koekebakker, S. (2005). Stochastic modelling of ﬁnancial electricity contracts. Dept. of Mathematics of University of Oslo, Pure Mathematics, no. 24, September. Chao, H.-P., Oren, S., and Wilson, R. (this volume). Reevaluation of vertical integration and unbundling in restructured electricity markets. Chapter 1. Correia, T.B. (2004). Modelo de Stackelberg na competição de empresas privadas e estatais pela expansão da oferta de energia elétrica. M.Sc. Dissertation – Unicamp. Correia, T.B., Melo, E., and Costa, A.M. (2006). Análise e avaliação teórica dos leilões de compra de energia elétrica proveniente de empreendimentos existentes no Brasil. Revista Economi A/ANPEC, 7(3). Correia, T.B., Melo, E., Silva, A.J., and Costa, A.M.A. (2005). Contra-Reforma institucional da indústria elétrica brasileira e novas perspectivas de mercado. VI Congresso de Regulação, Associação Brasileira das Agências de Regulação (ABAR), May 2005, Manaus. Correljé, A.F. and De Vries, L.J. (this volume). Hybrid electricity markets: the problem of explaining different patterns of restructuring. Chapter 2. Fernandez, V. (2002). The derivatives markets in Latin America with an emphasis on Chile. Working paper 36, Centro de Gestión de Universidad de Chile. Ford, A. (this volume). Global climate change and the near-term response of the electric power industry. Chapter 14. International Energy Agency (2003). Power Generation Investment in Electricity Markets. OECD/IEA. Joskow, P.L. (2006). Electricity sector liberalization: lessons learned from cross-country studies. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds), Elsevier. Joskow, P.L. and Schmalensee, R. (1983). Markets for Power. MIT Press. Klemperer, P. (2004). Auctions: Theory and Practice. Princeton University Press. Leal, C.F.C. (1998). Ágios, envelopes e surpresas: Uma visão geral da privatização das distribuidoras estaduais de energia elétrica. Revista do BNDES 12/1998, available at http://www.bndes.gov.br

EBL

572

Competitive Electricity Markets

Lee, W.W. (2004). US lessons for energy industry restructuring: Based on natural gas and California electricity incidences. Energ. Pol., 32, 237–59. Moran, A. and Skinner, B. (this volume). Resource adequacy and efﬁcient infrastructure investment. Chapter 11. Newbery, D. (2006). Electricity liberalisation in Britain and the evolution of market design. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds). Elsevier. Pinto Jr., H.Q. (1993). Financement, investissement et mode d’organisation des industries electriques: Le cas des pays d’Amérique Latine. Doctorate thesis in applied economics – Université Pierre Mendes-France de Grenoble, Centre National de la Recherche Scientiﬁque, Institut d’Economie et de Politique de l’Energie, Grenoble. Pires, J.C.L. and Piccinini, M. (1999). A Regulação dos setores de infra-estrutura no Brasil. In A Economia Brasileira nos Anos 90 (F. Giambiagi and M.M. Moreira, orgs).BNDES, 217–60. Pires, M.C.P. (2000). Regulação e concessão de serviços públicos de energia elétrica: Uma análise contratual. M.Sc. Dissertation, IE, UFRJ, Rio de Janeiro. Raineri, R. (2006) Chile: Where it all started. In Electricity Market Reform: An International Perspective (F.P. Sioshansi and W. Pfaffenberger, eds), Elsevier. Rodriguez-Pardina, M. and Estache, A. (1996). Exploring market-based options for a reformed Brazilian electricity sector. Working paper, The World Bank, Washington D.C. Sioshansi, F.P. (this volume). Electricity market reform: What have we learned? What have we gained? Introduction.

EBL

Index

Ackermann, T., 481–3 Active load management, 308, 368 Adib, P., 147, 252, 267, 274, 284, 290, 297, 299, 312, 318, 327, 384, 405, 550, 564 Adjustment auctions, 567 Administrative allocation rules, 166 Administrative demand curve, 355–6 Administrative scarcity, 364 Air-trafﬁc controllers, 142 Alanne, K., 472 Alexander, I., 70 Allegheny Power Company, 366 Alternative energy sources, incentives program for, 559 Alvarado, F., 155 American Transmission Company, 149 Amundsen, E., 6, 17, 552, 553 Anderson, K., 169 Anglo-Saxon models, 81, 83 Anthropogenic emissions, 502, 505, 506–509, 513, 514, 523, 539 Anti-competitive behavior, 272, 280, 284, 291 Antitrust law, 269 Arango, S., 77 Asano, H., 452 Assembly Bill, 32, 518, 519 Auction: designs, 185–8, 193, 194, 237, 239 market rules, 182 revenue rights, 165, 230, 366 Australia’s Electricity Code, 415 Australian Energy Market Commission, 402, 412 Australian Energy Regulator, 287, 335, 354, 396, 397 Australian National Electricity Market, 253, 388, 389, 410 Australian pool mechanism, 407 Automatic buffer, 37 Automatic mitigation procedure, 294 Averch–Johnson effect, 50

Bacon, R.W., 86 Balanced schedule requirement, 252, 317 Balancing energy service, 253, 279 Balancing mechanism, 14, 105, 117, 126–8, 132, 133, 388 Balancing up load program, 317 Baldick, R., 238, 494 Barbose, G., 297, 310 Barcelona European Council, 98 Barker, J. Jr., 37 Barrett, L., 310 Barroso, L.A., 337 Baughman, M., 303 Bauknecht, D., 21, 469, 480 Baumol, W.J., 274 Belmans, R., 101 Benth, F.E., 552 Bergström, M., 311 Bertram, G., 486 Besant-Jones, J., 86 Besnosik, R., 551 Bezerra, B., 358 Bid-ask method, 185 Bid-based auction markets, 180 Bilateral transactions, 187 Binational Itaipu plant, 559 Bird, L., 526 Black-start capacity, 72 Blumsack, S.A., 79, 80 Blumstein, C., 39 Boiteaux, M., 336 Boiteux, M.P., 388 Bolinger, M., 455, 465 Bonbright, J.C., 29 Bonneville power administration, 29 Borenstein, B., 281 Borenstein, S., 35, 37, 39 Bowen, M., 503, 504, 509 Brazilian electricity supply industry, 569 Brazilian hydropower system, 545 Brazilian power system, 545, 555, 560 Brazilian privatization program, 548 British electricity trading and transmission arrangement, 6, 246 Brunekreeft, G., 21, 83, 478, 481, 484, 485, 493

EBL

Backstop: call option, 359 technology, 537

573

574 Bulk: electricity market, 548, 557 market organism, 547 Burges, K., 474 Burns, P., 478 Bush, G.W., 520 Bushnell, J.B., 396 California crisis, 16, 37–9, 42, 44, 48, 51, 54–7, 63, 146, 147, 158, 193, 281 California energy commission, 356, 454 California independent system operator, 286, 292, 356 California public utilities commission, 27, 60, 405, 426, 454, 500, 519 Californian market model, 115 Call option obligations, 357 Cao, D.M., 475, 476, 486 Cap-and-trade, 517–22, 524, 528, 538, 539 Capacity deﬁciency charge, 367, 369, 373 Capacity-deﬁciency penalty, 372 Capacity emergency transfer limit, 172 Capacity market design, 377 bipolar nature, 405 fundamentals, 368 implementation of a, 367 structure, 371 Capacity mechanism, 66, 73, 74, 76, 78, 79, 87, 328, 334, 338, 343, 346, 347, 350, 357, 360 design of, 74 rejection of, 360 Capacity payments, 235, 333–7, 345, 346, 348, 350, 351, 355, 356, 360, 380, 388, 391, 405, 411 Capital costs (Capex), 60, 355, 409, 481, 482, 484, 553, 566 Caramanis, M. C., 388 Carbon cycle, 504 Carbon sequestration, 506, 533, 534, 537 Carbon tax, 389, 402, 408, 425, 499, 500, 518, 520–3, 529, 538, 539 Cardell, J., 473 Casazza, J., 146 Cavaliere, C., 74 Caves, D., 299 Central commitment model, 262 Central Electricity Generating Board, 84 Centralized resource adequacy market, 340 Centralized vs. decentralized markets, 246 Centrally committed markets, 253–6, 261 Cervigni, G., 312 Chamber of Electricity Trading (CCEE), 557 Chandler, A.D., 29

Index Chandley, J.D., 142 Chao, H-P., 13, 27, 37, 63, 70–2, 165, 570 Chen, M., 36 Cherry-picking, 490 Chicago climate exchange, 524 Claussen, M., 507 Clean development mechanism, 515 Climate change program, 528 Climate Stewardship Act, 500, 517, 531 Coal: -based generation, 399 exploitation, 87 -ﬁred plants, 401 Coll-Mayor, D., 476 Combined cycle, 61, 341, 342, 350, 375, 527, 535, 536 Combined-cycle gas turbines, 54, 247, 255 Combined heat and power, 21, 426, 439, 469 Combustion turbine, 42, 219, 341–3, 375, 378 Commission for the regulation of energy, 287 Commonwealth Edison Company Control Area, 366 Competitive benchmarks, 16, 245, 246 Competitive electricity markets, regulator in, 9 Conduct-impact test, 231, 232 Congestion management, 15, 50, 72, 76, 101, 105, 110, 114–19, 122, 123, 128, 133, 135, 136, 141–4, 158, 180, 185, 199, 202, 214, 236, 284, 312 guidelines for, 112, 122 Congestion metrics, 142, 144, 162, 164, 226 Congestion rents, 161, 162, 168 Connor, P.M., 480–2 Consumer surplus, 192, 240 Contract for differences, 44 Contracting obligations, 358 Contractual remedies, 231 Controllable load, 313 Cook, G., 310 Cooper, R., 521, 525 Correia, T.B., 22, 559, 570 Correljé, A.F., 14, 28, 65, 68, 70, 74, 87, 335, 545, 550 Cost allocation for transmission, 174 Cost-causation, principle of, 221 Cost evolution of technologies, 464 Cost of new entry, 344, 378, 379 Cost-of-service regulation, 29–34, 45, 60–3, 161 Council of European Energy Regulators, 112 Cox, P.M., 511 Crampes, C., 84 Cramton, P., 19, 190, 328, 333, 342, 358, 364, 382, 388 Cronin, F.J., 79

EBL

Index Cross-border: balancing services, 127, 128 investment, 121, 125 tariffs, 112 trade task force, 118 trade, distortions of, 114 transport of power, 98 Cross sound cable, 172, 173 Cross-subsidies, 11, 13, 31, 34, 188, 408, 561 Cumulative price cap, 403 Cumulative price threshold, 354 Curious enigma, 506 Currency devaluation, 5 Curtailment programs, 304 Danish grid, 457 Danish wind association, 458 Day-ahead market, 42, 108, 143, 144, 147, 161, 179–88, 191, 194–226, 234, 237, 252, 262, 305, 312, 313, 317, 320, 356, 370, 377, 393 De Araújo, J.L.R.H., 6, 22, 78, 543–5, 547–54, 556, 557, 559, 562, 565, 568–70 de Jong, J., 486 De Vries, L.J., 14, 28, 65, 68, 73, 74, 87, 335, 486, 545, 550 Dehdashti, E.S., 66 Deliverability issues, 359 Delivered price test, 278 Demand bids, 195 Demand charge, 47, 230, 336 Demand curve, 189, 194, 199, 208, 209, 240, 302, 321, 340, 342–5, 355, 356, 363, 377–80 Demand participation, 14, 17, 139, 297 Demand response coordinating council, 298 Demand-side resources, 313, 318, 353 Demsetz, H., 70 Design deﬁciencies, 294 Devine, W.D. Jr., 29 Differences settlement price, 558 Distributed generation, 21, 469, 470, 472, 475, 559 Distribution grid, deﬁnition of, 470 Distribution network operators, 470 Diversiﬁcation of generation, 58 Don Corleone Equilibrium, 268 Dual fuel, 394 Duquesne light company, 366

575 Effective market monitoring, 282 Ehlers, E., 481, 484, 485, 493 Electric Reliability Council of Texas, 154, 272, 300, 311, 327 Electric reliability organization, 144, 146, 153 Electricity auction design, 183, 185 Electricity distribution companies, 374 Electricity focus group, 118 Electricity network regulation, 478 Electricity pool, 246, 252–4, 261 Electricity regional initiative, 118 Electricity supply industry, 1, 22, 67, 73, 469, 543, 550 Electricity task force, 121 Ellerman, A.D., 522–5 Emergency curtailment program, 317 Emergency demand response program, 307 Emergency interruptible load program, 319 Emergency load response, 308 Emissions reduction: goals for, 514 incentive for, 521 Emissions trading scheme, 22, 500, 524 Energy contracting procedures, 558 Energy efﬁciency appliances, 349 Energy information administration, 529, 531, 532, 534, 537, 538 Energy intermarket surveillance group, 273 Energy market rules, 204, 217 Energy Policy Act, 3, 9, 33, 141, 146, 147, 298, 306 elements of, 175 Energy services companies, 301 England–Wales Pool, 37, 38 Enron bankruptcy, 148 Equilibrium climate sensitivity, 509, 510 Estache, A., 562 European electricity markets, 81, 103, 134 European emissions trading scheme, 524 European federation of energy traders, 104, 129 European grid, 96, 100, 132 European power exchanges, 108 European regulator group for electricity and gas, 118 European renewable energy council, 420 European transmission system operator, 100, 142 European Union, 2, 3, 41, 66, 95, 96, 100, 125, 192, 420, 500, 517, 552 EuroPEX, 112, 123

EBL

East central area coordination agreement, 152 Economic: crises, 3, 78 slump, 5 withholding, 231, 277, 293, 370 Economically disadvantaged customers, 62

Fabra, N., 84, 189 Fabrizio, K. M., 396 Fairﬁeld, H., 526, 529

Index

576 Falk, J., 274 Faruqui, A., 298, 299 Federal Energy Regulatory Commission, 7, 8, 32, 53, 79, 80, 87, 116, 176, 180, 191, 221, 285, 289, 291, 298, 327, 338, 348 jurisdiction, 54 regulations, 86 Federal Network Agency, 492 Federal Power Act, 146, 147, 153, 167, 174, 192 Feed-in tariffs, 21, 421, 424, 428, 438, 442 Felder, F., 172 Fens, T., 71 Fernandez, V., 552 Fertilization effect, 511 Fiddaman, T., 522, 523 Financial transmission rights (FTRs), 116, 143, 155–7, 159–62, 164–8, 172, 197, 198, 225–7, 229, 230, 235, 236, 239, 365 payment deﬁciencies, 230 Finon, D., 67, 68, 70, 72, 83, 84, 87, 419, 463 Flannery, T., 502, 503 Flavin, Ch., 420 Florence Forum, 110, 112, 119–24, 128, 129 Flowgates, 164, 165 prices, 226, 241 Folk, S., 310 Forced outage rate, 364, 380 Ford, A., 21, 419, 499, 510, 519, 520, 529, 534–8, 570 Forecast required energy for dispatch, 212 Foreman-Peck, J., 70 Forest, C., 513 Forthcoming auction footing, 394 Forward capacity market, 334, 338, 345 Fouquet, D., 520 Free contracting environment, 555, 557 Friedman, M., 70 Fringe competition, 84 Fuel-intensive technologies, 551 Fuel-switching, 528, 536, 539 Full funding of the rights, 230 Functional unbundling, 58, 96, 113, 121, 130, 132 Future nodal market, 319

Glachant, J.-M., 67, 68, 70, 83, 84, 87, 112, 136, 137 Global climate change, 21, 499 Global warming, 402, 408, 499, 501–505, 512, 513, 519, 539 Goldman, C., 297, 311 Gore, A., 503, 521 Goto, M., 4 Government-sanctioned monopolies, 29 Govindasamy, B., 511 Green certiﬁcates, 53, 429, 463, 464 Green electricity can, 440 Green power, 57, 407, 433, 450, 526 Green, J., 254 Green, R., 85, 86 Greenhouse Effect, 503 Greenhouse gases, 21, 22, 53, 423, 427, 499–500, 503, 506, 512–14, 517–20, 529 accumulation of, 21 Grey, P., 406 Grid and market control, 36 Grid congestions, 485 Grid management, 37–9, 43, 49, 50 charges, 43, 49, 150 Guan, X., 237, 249 Gumerman, E.Z., 474, 484

EBL

Gas-fueled combined cycle, 527, 532, 534, 537 Generation and transmission, Centralized dispatch of, 30 Generation-owning utilities, 367 Generator bidding freedom, 403 Generator ramp rates, 50, 189 George, S. S., 299 German rooftop PV programs, 453 Gill, R.A., 506

Haaland, 85 Haas, R., 6, 7, 66, 82, 85, 424, 428, 453, 465, 466, 472, 473, 554 Hallett, I., 318 Hampton Court, 99 Hancher, Lee, 112 Hansen, J., 502, 508, 511 Harmon, R., 526 Harris, C., 70 Harvey, S., 159 Haydon, J.J., 411 Heating and cooling applications, 35 Held, A., 21, 419, 448 Heller, T.C., 66, 67, 71, 73, 79, 86 Helm, D., 70 Helman, U., 7, 15, 179, 192 Henisz, W.J., 78 Henney, A., 142 Herﬁndahl–Hirschman Index (HHI), 278 High Voltage DC (HVDC) technology, 169, 172 Hill, J., 527 Hirschhausen, C., 478 Hirschhorn, M., 526 Hirst, E., 169, 170, 300 Hobbs, B.F., 15, 179, 186, 193, 197, 237, 249, 343 Hogan, W. W., 73, 142, 159, 164, 166, 221, 226, 238, 247, 248, 250, 333, 358, 383, 411

Index Holt, D., 480 Hortaçsu, A., 252 Houghton J., 505, 510 Houston Lighting and Power Company, 318 Hrab, Roy, 6 Huber, C., 466 Hunt, S., 32, 37, 71, 80, 247 Hurlbut, D., 16, 267, 351 Hybrid electricity markets, 13, 65, 77 Hydroelectric dams, 29 Iannucci, J.J., 474, 484 IMF, 70, 86, 87 Impact tests, 280 Incentive-based demand response, 305 Increment bids, 224 Independent market monitor, 273, 282 Independent power producers, 29, 66, 146, 545, 556 Independent system operator, 7, 15, 16, 36, 38, 39, 44, 48, 50, 53, 81, 132, 142, 164, 179–86, 189, 191, 193, 195, 198, 199, 201, 209, 210, 218, 220, 224, 226, 238, 275, 283, 286, 288, 291, 298, 311, 328, 344, 345, 548 Inframarginal proﬁt, 271, 275, 276, 331, 334 Installed capacity, 307, 337 Institute of Public Affairs, 387, 390 Integrated global systems model, 500, 507 Integrated resource planning, 49, 54, 144, 161, 171, 343 Inter-temporal smoothing of rates, 62 Inter-TSO compensation, 110, 117 Inter-zonal congestion, 162 Intergovernmental Panel on Climate Change, 500–504, 539 International Energy Agency, 440, 476, 552 International Transmission Company, 149, 171 Intra-communal barriers, 85 Intra-zonal: congestion, 40, 158, 159, 162, 163 transmission capacity, 40 Investment-based tax incentives, 455, 456 Investment risks, Allocation of, 1 Investor-owned utilities, 29, 268, 312 Irrational utilization of capacity, 135 Isser, S., 308

577 Joskow, P.L., 4, 5, 8, 30, 54, 57, 65–7, 70, 79, 80, 86, 169, 247, 333, 382, 388, 405, 482, 491, 492, 543, 550, 570 Kahn, A.P., 189, 271 Kahn, E., 238, 494 Kerr, R., 502, 511 Kimura, O., 452 Kinnunen, K., 84 Kirchhoff’s current and voltage laws, 241 Klemperer, P., 185, 189, 564 Knops, H.P.A., 72 Koekebakker, S., 552 Koenig, S.A., 249 Komor, P., 420 Koszalka, M., 310 Krein, S., 315 Krishna, V., 185–7, 189 Kump, L.R., 511 Künneke, R., 71 Kwoka, John, 66, 74 Kyoto protocol, 515, 524, 527 Lagrangian relaxation, 236, 237, 249 Langniss, O., 449, 463 Leal, C.F.C., 562 Lee, W.W., 564 Lekander, P., 528 Lemming, J., 463 Lenssen, N., 420 Leveque, F., 273 Lévêque, F., 70 Littlechild, S., 11, 70, 406 Litvinov, E., 341, 342 LNG, 77 Load-duration proﬁle, 30, 37, 47, 58, 59 Load serving entities, 165, 328, 336, 368 Load-serving entities, 41, 167, 226, 317, 355 Load shifting, 298 Local market power, 274, 293 Locational marginal pricing (LMP), 16, 79, 179, 181, 188–91, 200, 213, 224, 237, 238, 255, 302, 359, 375 features of, 158 principles of, 179 Lohr, S., 529 Lomborg, B., 513 Loop ﬂows, 236 Lopez, A., 453, 454, 465, 466 Losekann, L., 556 Louisville Gas and Electric and Kentucky Utilities, 150 Lund, P., 472, 478 Luoma, J., 510

EBL

Jacoby, H.D., 513, 522 Jamasb, T., 66, 69, 74 Jaussaud, D., 267 Jenkins, N., 471, 473, 487 Johnson, R.B., 237, 250, 251 Jones, C.D., 511 Jones, S., 318

Index

578 McDaniel, T., 478 McNamara, R., 340 Macro-economic characteristics, 68, 89 Macro-economic crisis, 77 Marginal congestion: computation of, 225 losses, settlement of, 224 Marginal cost: based offers, 337 estimation of, 553 Market-based economy, 86 Market-based tariffs, 78 Market-clearing price (MC), 180, 181, 190, 219, 231, 232, 270, 271, 275–80, 283, 313, 314, 331, 337, 351, 377, 383, 522, 527 Market-clearing processes, 53 Market for regulation and reserves, 199, 207, 214, 218 Market harmonization, 121 Market manipulation, 291, 293 Market monitoring, 16, 148, 268, 281, 282, 284–7 Market-oriented political culture, 436 Market performance, 14, 139, 402 Market power abuse, 293, 351 Market pricing, 197, 213 Market reform evolution, 13, 25 Market simulations, 258 Market surveillance administrator, 286, 287, 292 Martinot, E., 439 Mas-Colell, A., 254 Mastrandrea, M.D., 512 Matláry, J., 85 Maximum emergency generation, 209, 232 Meeus, L., 101 Menanteau, P., 424 Mendelsohn, R.O., 513 Merchant transmission model, 172, 176 Meritet, Sophie, 83 Meyer, A., 526 Meyer, N.I., 21, 419, 420, 426, 457 Michaels R.J., 27, 411, 412 Michigan Electric Transmission Company, 149 Mid-Atlantic Area Council, 152, 310, 311 Middle Eastern gas suppliers, 85 Midwest independent system operator, 286, 287 Milward, R., 70 Ministry of Mines and Energy, 543, 555 Missing money, 19, 327–9, 331, 333–5, 355, 358, 387, 566 Mitchell, C., 442, 480–2 Mitigation measures, 231 Moran, A., 6, 19, 267, 328, 350, 387, 478, 550 Morgenstern, R.D., 522 Moskovitz, D., 482

Moss, D., 269, 270 Motluk, S.A., 79 Muckstadt, J.A., 249 Multi-unit auction design, 188 Multinational conglomerates, 85

National academy study, 510 National agency for electric energy, 555 National allocation plans, 525 National Association of Regulatory Utility Commission, 298 National Commission on Energy Policy, 500, 522 National Competition Policy, 390 National Council for Energy Policy, 555 National Economic and Social Development Bank, 570 National Electricity Code Administration, 287 National Grid Company, 14, 39, 142 National interest electric transmission, 174, 175 National system operator, 546, 549, 555 Natural Gas Bill, 570 Neenan, B., 301, 310 Neoclassic economic theory, 66 Neoclassical approach, 69 Network regulation, 470, 480 Neumann, S., 302 New electricity trading arrangement, 6, 39, 42, 44, 48, 51, 246, 396, 544, 551 New South Wales, 2, 11, 397–9, 402, 410 Newbery, D.M., 1, 6, 39, 65–7, 70, 71, 80, 252, 253, 396, 544 Nishio, K., 21, 419, 452 No-load offer costs, 190 Nodal pricing, see Locational marginal pricing (LMP) Nodal protocols, 320 Non-diversiﬁable risks, 63 Non-domestic customers, 101, 105, 113 Non-fossil fuel obligation, 441 Non-performance penalty scheme, 358 Non-spinning reserves, 37, 310, 313, 315, 317 Nordic market, 6, 17, 553 NordPool, 37–41, 43, 44, 48, 51, 54, 102–109, 253, 285, 286, 332, 336, 425, 544, 551–3 Normative value, 561 North, D., 67, 68, 74, 75, 88 Northeast power coordinating council, 154, 310, 311 Northern Hemisphere, 503 Not-willing-to-pay congestion, 214

EBL

Index O’Neill, R.P., 7, 15, 41, 165, 172, 179, 180, 182, 188, 191, 192, 238, 239, 245, 249, 258, 259, 261 Occasional high-price excursions, 414 OECD countries, 66–8, 70, 77, 79, 81 Off-peak period, 211, 215, 224 Oligopolistic industry, 74 Oliveira, A., 562 Oosterhuis, F., 420 Operating reserves charges, 224 Operational costs (Opex), 230, 482, 484, 560 Order, 15, 141, 144, 147, 155–7, 164, 166, 176, 888 Oren, S.S., 13, 16, 18, 27, 245, 250, 274, 302, 318, 327, 331, 333, 357, 384, 388, 405, 406 Organizational unbundling of ﬁrms, 28 Out of market transactions, 43, 191 Out-of-merit generation, 158, 160, 162, 232, 234 Output gap analysis, 279 Over-the-counter (OTC) markets, 105, 551 Paltsev, S., 523–5, 528 Pan-European electricity market, 3, 85, 98 Pan-European energy regulators, 112 Pan-European technical agencies, 109 Pancaking of transmission charges, 36 Parker, M.J., 70, 85 Patton, D., 319 Pay-as-bid pricing, 181, 189, 190, 271 Peak and offpeak periods, 47, 57 Peak clipping, 298, 318 Peck, S., 165 Pennsylvania-New Jersey–Maryland Interconnection (PJM), 16, 18, 80, 165, 167, 182, 195, 207, 209, 213, 219, 237, 238, 291, 328 emergence of a capacity market in, 366 management, 289 pivotal supplier test, 278, 380 Pennsylvania Public Utility Commission, 367, 368, 373 Pepermans, G., 472 Perez, Y., 463 Performance-based regulation of utilities, 63 Pervasive market power, 272 Petersen, E.L., 457 Pfaffenberger, W., 2–5, 9, 10, 67 Photovoltaics, 420 Piccinini, M., 563 Pinto Jr., H.Q., 562 Pires, J.C.L., 563 Pires, M.C.P., 560 Pizer, W., 521, 522 Policies to reduce emissions, 520 Pollitt, M.G., 66, 396 Pool selling price, 442

579 Poolco model, 247–9, 252–3 Portney, P., 521 Post-communist institutional framework, 83 Potts Voll, S., 86 Power exchanges, 34, 39, 44, 71, 104, 106, 107, 117, 180 Power procurement: agreements, 54 risks, 301 Powernext, 102, 105–108 Prediction and policy analysis, 500, 527 Prevention of harmful market activities, 282 Price: deregulation, 96 mechanism, 78 regulation, 65, 482 -sensitive demand bids, 209, 212 war, 412 Prinn, R., 507 Priority interconnection plan, 99 Privatization and liberalization scheme, 1 Producer surplus, 192, 197, 240 Production tax credit, 434, 456, 520, 523, 537, 538 Proﬁt-maximization problem, 263–5 Public monopolies, 72 Public Utilities Regulatory Policies Act, 32, 146, 426, 439 Public Utility Commission of Texas, 270, 272, 292, 312, 319, 327, 330, 350–2 Public Utility Holding Company Act, 144, 175, 176

EBL

Qualiﬁed scheduling entities (QSEs), 316 Qualifying facilities, 32, 147, 330, 426, 439 Quasi-independent regulator, 84 Quota-based trading, 421, 444 Ragwitz, M., 423, 436, 446, 456, 462 Raineri, R., 1, 6, 19, 544 Rajaraman, R., 155 Ramp rate, 36–8, 42, 43, 48, 54, 181, 187, 195, 197, 208, 212, 219, 234, 241, 255, 262 Rate-of-return regulation, 50, 367, 482–4 Rate shocks, 9 Raven, R., 472 Real-time auction pricing rules, 191 Real-time buyers, 229 Real-time congestion management, 214 Real-time market, 40, 51, 144, 157, 180–2, 187–9, 191, 194, 196, 198, 199, 207, 208, 210–15, 217–19, 221, 224, 234, 238, 282, 305, 308, 335, 356 offer period, 217

Index

580 Real-time pricing, 47, 213, 297, 304, 306, 310, 349 Regional electricity markets task Force, 118 Regional energy markets, 119 Regional harmonization, 114 Regional transmission organization, 15, 36, 80, 100, 141, 142, 147, 236, 238, 285, 305 Regulated contracting environment, 555, 557 Regulatory: barriers, 303 function, 73 Reilly, J., 507, 523–5, 528 Reliability assurance agreement, 368, 369, 370 Reliability must run, deployment of, 339 Reliability Pricing Model, 18, 172, 335, 338, 343, 363, 379 Reliability unit commitment, 182, 191, 195, 210–12, 215–17, 221, 223, 238, 239, 255, 320, 333 Renewable Energy Act, 430, 431, 439, 440 Renewable energy resources, 21, 54, 112, 419, 420, 423, 427, 429, 439, 451 electricity from, 419 Renewable energy schemes in Japan, 451 Renewable generating technologies, 539 Renewable portfolio standards, 20, 21, 425, 449, 520 Reserve trader, 404, 414 Residual loss payment, 230 Resource adequacy, 5, 17–19, 325, 327–9, 340, 351, 354, 356, 387 Restructured markets, 17, 297, 304 Restructuring policy, 67, 70, 83 Retail energy providers, 56, 355 Retail liberalization, 52 Retail price regulation, 79 Return on equity, 174, 175 Revenue sufﬁciency guarantee, 16, 182, 198, 211, 212, 214, 220, 221, 224 Riechmann, C., 478 Risk management, 11, 28, 29, 41, 58, 60, 63, 147, 299, 330, 333–5, 348, 357, 358, 360, 395, 405, 512 Roberts, V., 476 Rodriguez-Pardina, M., 562 Roques, F.A., 388 Rosenzweig, M. B., 86, 299 Ross, A.M., 250 Rotger, J., 172 Rudkevich, A., 159 Rudnick, H., 37, 65, 67, 77, 78, 86 Ruff, L.E., 142, 248

Sarmiento, J.L., 506 Scarcity pricing, 16, 182, 183, 189, 192–4, 199, 209, 210, 214, 231, 328, 333, 337, 339, 347, 348, 351–6, 363, 364, 377, 382–4 lack of, 338, 339 versus high prices, 270 Scheer, H., 420 Scherer, F.M., 69 Schmalensee, R., 54, 247, 491, 543 Schneider, S.H., 503, 512 Schubert, E., 18, 274, 281, 318, 327, 348, 351, 405 Schweppe, F. C., 197, 200, 225, 227, 240, 303 Sea breeze proposal, 173 Security-constrained unit commitment, 179, 195, 204, 210 Self-committed market, 16, 245, 246, 253, 254, 256, 257, 259, 261, 262 Senate Bill, 139, 500, 517 Serrallés, R.J., 84 Seven brothers, 85 Seven “wedges”, 515 Shanker, R.J., 328 Shepherd, W.G., 274 Sherman, R., 50 Shuttleworth, G., 66, 71 Siddiqui, A.S., 165 Simshauser, P., 388 Singh, H., 15, 141, 150, 302 Sioshansi, F.P., 1–4, 9, 10, 16, 65, 67, 70, 86, 297, 543 Sioshansi, R., 245, 251 Skinner, B., 6, 19, 267, 328, 350, 387, 550 Smart meters, 310, 408 Socolow, R., 506, 516 Solid biomass, 420 Southern California Edison Company, 308 Southwest Power Pool, 8, 144, 154, 285, 286, 311, 332 Spark-spread contract, 55 Special case resources program, 307 Spine, H.N., 301 Spot energy auctions, 237 Spot markets, 42 Stand-alone transmission companies, 149, 169, 175 Standard market design, 38, 79, 116, 146, 147, 191, 236, 246, 248, 285, 307, 328, 340 Sterman, J.D., 506 Sterzinger, G., 523 Stiglitz, J., 69 Stoft, S., 19, 65, 73, 182, 188, 190, 191, 194, 239, 269, 328, 333, 342, 358, 364, 382, 387, 388

EBL

Saari, A., 472 Samuelsohn, D., 526, 529

Index

581 UK Power Exchange, 38 Unbundling, 13, 27, 38, 76, 82, 84, 88, 129, 390, 481 monopolies, 390 networks, 72, 470 supplies, 38 US Acid Rain Program, 525, 526, 528 US Department of Justice, 269 US National Energy Policy, 411 US transmission system, 169 Utilities cost recovery, 61

Strategic European energy review, 99 Strbac, G., 476, 477 Streiffert, D., 237, 255 Sussex Energy Group, 476 Suzuki, T., 452 Sweeney, J.L., 6, 39, 147, 188, 192, 193, 281 Sweeney, L.B., 506 Swisher, J.N., 472, 473 Synchronous reserve, 184, 208, 218 Systemic risks, 59 Tabors, R., 473 Takahashi, K., 487 Tariff adjustments, 78 Tariffs for geothermal electricity, 440 Temporal market power, 274 Tennessee Valley authority, 29 Thermal plants, priority program for, 549, 563 Third-party demand reduction, 301 Thomas, S., 66, 85, 86, 487 Three pivotal supplier test, 233, 380 Tiered frequency response program, 319 Tirole, J., 382 Tol, R., 513 Top-down harmonization, 135 Top-down resource adequacy mechanisms, 329, 335 Tradeable green certiﬁcates, 520 Trading chamber for electric energy, 555 Trans-European energy networks, 99 Trans-European Networks, 125 Transaction costs, 494 Transmission: assets, 132, 142, 148–50, 156, 176 congestion, 11, 31, 36, 38, 42, 51, 158, 161, 164, 169, 189, 226, 232, 280, 294, 303, 313, 359, 411 expansion, process of, 160 investment, 15, 32, 49, 50, 59, 118, 125, 141, 144, 150, 161, 162, 169–76, 299, 303, 381, 389, 411, 559 loading relief procedures, 155 network manager, 72 organizations, Evolution of, 145 system operators, 14, 95, 142, 473 tariff harmonization, 123 upgrade decisions, 161 usage, 198, 238 Treblicock, Michael, 6 Trefny, F., 317 Tschamler, Taff, 9, 10 Twele, J., 474 Twelemann, S., 83

Value of lost load, 194, 333, 388, 404 Van Damme, E., 84 Van der Linde, J.G., 87 van Overbeeke, F., 476 Variable resource requirement, 335, 343, 344, 377 Varian, H., 513 Varming, S., 476 Vázquez, C., 357 Verbruggen, A., 449 Verhaegen, K., 101 Vertical demand function, 343, 344 Vertical integration: argument, 27, 63 historical motives, 29 Vesting contracts, 45, 54, 391 Vickrey–Clarke–Grove mechanisms, 193, 254 Victor, D.G., 65–7, 71, 79, 86 Virtual demand bids, 200 Virtual interconnection arguments, 176 Vognild, I.H., 311 Volume-related revenue driver, 490 Volumetric insurance, 33 Voluntary mitigation plans, 352 Voogt, M.H., 420 Vostock ice core, 503

EBL

Wallenrod, M., 309 Walrasian auction model, 257 Walther, B., 311 Weart, S.R., 503, 510, 511 Webster, M., 507–509, 512, 513, 516 Weinmann, J., 68, 74, 81 Western area power authorities, 29 Western electricity coordinating council, 154, 310, 500, 533 Weyant, J., 527 Whinston, M., 254 Wholesale: electricity markets, 73, 268 generators, 147 Williamson, O.E., 30, 67, 68, 74, 75

582 Willman, P., 74 Wilson, R., 13, 39, 47, 182, 191, 250, 251, 257 Wind energy in Denmark, 456 Windfall proﬁts, 448, 449, 459, 462, 526 Winner determination rule, 263 Wiser, R., 422, 436, 439, 442, 447–9, 451, 455, 460, 461, 463, 465, 466 Wolak, F., 281 Wolfram, C., 396 Wollenberg, B., 159 Wolsey, L.A., 251 Wood, A.J., 159 World alliance for decentralized energy, 471 World Bank, 70, 85, 86, 87, 513, 524

Index Yajima, M., 4 Yang, J., 341, 342 Yi-chong, X., 86

Z-factor, 489 Zarnikau, J., 17, 47, 147, 252, 253, 267, 297, 298, 312, 313, 315, 318, 320, 348, 349, 353, 354, 473 Zelner, B.A., 78 Zonal locational pricing, 198 Zonal pricing of transmission, 39, 40 Zone-based transmission, 116

EBL

Competitive Electricity Markets: Design, Implementation, Performance (Elsevier Global Energy Policy and Economics Series)

Reshaping European Gas and Electricity Industries (Elsevier Global Energy Policy and Economics Series)

European Electricity Systems in Transition (Elsevier Global Energy Policy and Economics Series)

European Electricity Systems in Transition (Elsevier Global Energy Policy and Economics Series)

Competitive Electricity Markets And Sustainability

European Energy Industry Business Strategies (Elsevier Global Energy Policy and Economics Series)

National Reforms in European Gas (Elsevier Global Energy Policy and Economics Series)

Electricity Market Reforms: Economics and Policy Challenges

Electricity markets

Energy Economics: Concepts, Issues, Markets and Governance

International Economics: Global Markets and International Competition

Green Energy: Technology, Economics and Policy

Green Energy: Technology, Economics and Policy

Quantitative and Empirical Analysis of Energy Markets (World Scientific Series on Energy and Resource Economics)

Electricity (Early Bird Energy)

Energy Market Experience Learning from the Blackouts: Transmission System Security in Competitive Electricity Markets

Energy Sector Management Managing sustainability in developing electricity markets

Power Generation Investment in Electricity Markets (Energy Market Reform)

Managing Enterprise Risk: What the Electric Industry Experience Implies for Contemporary Business (Elsevier Global Energy Policy and Economics)

Wind Energy in the 21st Century: Economics, Policy, Technology and the Changing Electricity Industry

BGP Design and Implementation

Economics And Energy

BGP Design and Implementation

Renewable Energy: A Global Review of Technologies, Policies and Markets

Global Competitive Strategy

Load and Global Response of Ships (Elsevier Ocean Engineering Series)

Energy Balances and Electricity Profiles 2005

Cannabis Policy, Implementation And Outcomes

Football Economics and Policy

Energy Pricing: Economics and Principles (Energy Systems)

Load and Global Response of Ships (Elsevier Ocean Engineering Series)

Competitive Electricity Markets: Design, Implementation, Performance (Elsevier Global Energy Policy and Economics Series)

Reshaping European Gas and Electricity Industries (Elsevier Global Energy Policy and Economics Series)

European Electricity Systems in Transition (Elsevier Global Energy Policy and Economics Series)

European Electricity Systems in Transition (Elsevier Global Energy Policy and Economics Series)

Competitive Electricity Markets And Sustainability

European Energy Industry Business Strategies (Elsevier Global Energy Policy and Economics Series)

National Reforms in European Gas (Elsevier Global Energy Policy and Economics Series)

Electricity Market Reforms: Economics and Policy Challenges

Electricity markets

Energy Economics: Concepts, Issues, Markets and Governance

International Economics: Global Markets and International Competition

Green Energy: Technology, Economics and Policy

Green Energy: Technology, Economics and Policy

Quantitative and Empirical Analysis of Energy Markets (World Scientific Series on Energy and Resource Economics)

Electricity (Early Bird Energy)

Energy Market Experience Learning from the Blackouts: Transmission System Security in Competitive Electricity Markets

Energy Sector Management Managing sustainability in developing electricity markets

Power Generation Investment in Electricity Markets (Energy Market Reform)

Managing Enterprise Risk: What the Electric Industry Experience Implies for Contemporary Business (Elsevier Global Energy Policy and Economics)

Wind Energy in the 21st Century: Economics, Policy, Technology and the Changing Electricity Industry

BGP Design and Implementation

Economics And Energy

BGP Design and Implementation

Renewable Energy: A Global Review of Technologies, Policies and Markets

Global Competitive Strategy

Load and Global Response of Ships (Elsevier Ocean Engineering Series)

Energy Balances and Electricity Profiles 2005

Cannabis Policy, Implementation And Outcomes

Football Economics and Policy

Energy Pricing: Economics and Principles (Energy Systems)

Load and Global Response of Ships (Elsevier Ocean Engineering Series)

Recommend Documents